
# Импорт необходимых библиотек
import os
import optuna
import numpy as np
import pandas as pd
import seaborn as sns
import plotly.express as px
import matplotlib.pyplot as plt

from time import time
from sklearn.svm import SVR
from lightgbm import LGBMRegressor
from catboost import CatBoostRegressor
from sklearn.dummy import DummyRegressor
from sklearn.pipeline import make_pipeline
from sklearn.preprocessing import StandardScaler
from sklearn.linear_model import LinearRegression
from sklearn.ensemble import RandomForestRegressor
from sklearn.model_selection import train_test_split
from sklearn.metrics import mean_absolute_error as mae
from sklearn.model_selection import KFold, cross_val_score


# Обозначим константы
RANDOM_STATE = 140823


# Список локальных файлов
local_files = [
    'data_arc_new.csv',
    'data_bulk_new.csv',
    'data_bulk_time_new.csv',
    'data_gas_new.csv',
    'data_temp_new.csv',
    'data_wire_new.csv',
    'data_wire_time_new.csv'
]
# Список онлайн файлов
online_files = [
    'https://***/data_arc_new.csv',
    'https://***/data_bulk_new.csv',
    'https://***/data_bulk_time_new.csv',
    'https://***/data_gas_new.csv',
    'https://***/data_temp_new.csv',
    'https://***/data_wire_new.csv',
    'https://***/data_wire_time_new.csv'
]
# Создание пустого словаря для хранения наборов данных
datasets = {}
# Цикл по спискам локальных и онлайн файлов
for i in range(len(local_files)):
    local_file = local_files[i]
    online_file = online_files[i]
# Извлечение имени набора данных из имени файла
    dataset_name = local_file.split('.')[0]
# Проверка наличия локального файла
    if os.path.exists(local_file):
        # Если локальный файл существует, считываем его и сохраняем в словаре datasets
        datasets[dataset_name] = pd.read_csv(local_file)
    else:
        # Если локальный файл не существует, считываем онлайн файл по URL и сохраняем в словаре datasets
        datasets[dataset_name] = pd.read_csv(online_file)
# Отображение набора данных и информации о нем
    print(f"{dataset_name}:")
    display(datasets[dataset_name])
    print(datasets[dataset_name].info())
    print('*'*100)
# Присвоение отдельных наборов данных переменным для удобства
df_arc = datasets['data_arc_new']
df_bulk = datasets['data_bulk_new']
df_bulk_time = datasets['data_bulk_time_new']
df_gas = datasets['data_gas_new']
df_temp = datasets['data_temp_new']
df_wire = datasets['data_wire_new']
df_wire_time = datasets['data_wire_time_new']

data_arc_new:

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 14876 entries, 0 to 14875
Data columns (total 5 columns):
 #   Column                Non-Null Count  Dtype  
---  ------                --------------  -----  
 0   key                   14876 non-null  int64  
 1   Начало нагрева дугой  14876 non-null  object 
 2   Конец нагрева дугой   14876 non-null  object 
 3   Активная мощность     14876 non-null  float64
 4   Реактивная мощность   14876 non-null  float64
dtypes: float64(2), int64(1), object(2)
memory usage: 581.2+ KB
None
****************************************************************************************************
data_bulk_new:

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 3129 entries, 0 to 3128
Data columns (total 16 columns):
 #   Column   Non-Null Count  Dtype  
---  ------   --------------  -----  
 0   key      3129 non-null   int64  
 1   Bulk 1   252 non-null    float64
 2   Bulk 2   22 non-null     float64
 3   Bulk 3   1298 non-null   float64
 4   Bulk 4   1014 non-null   float64
 5   Bulk 5   77 non-null     float64
 6   Bulk 6   576 non-null    float64
 7   Bulk 7   25 non-null     float64
 8   Bulk 8   1 non-null      float64
 9   Bulk 9   19 non-null     float64
 10  Bulk 10  176 non-null    float64
 11  Bulk 11  177 non-null    float64
 12  Bulk 12  2450 non-null   float64
 13  Bulk 13  18 non-null     float64
 14  Bulk 14  2806 non-null   float64
 15  Bulk 15  2248 non-null   float64
dtypes: float64(15), int64(1)
memory usage: 391.2 KB
None
****************************************************************************************************
data_bulk_time_new:

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 3129 entries, 0 to 3128
Data columns (total 16 columns):
 #   Column   Non-Null Count  Dtype 
---  ------   --------------  ----- 
 0   key      3129 non-null   int64 
 1   Bulk 1   252 non-null    object
 2   Bulk 2   22 non-null     object
 3   Bulk 3   1298 non-null   object
 4   Bulk 4   1014 non-null   object
 5   Bulk 5   77 non-null     object
 6   Bulk 6   576 non-null    object
 7   Bulk 7   25 non-null     object
 8   Bulk 8   1 non-null      object
 9   Bulk 9   19 non-null     object
 10  Bulk 10  176 non-null    object
 11  Bulk 11  177 non-null    object
 12  Bulk 12  2450 non-null   object
 13  Bulk 13  18 non-null     object
 14  Bulk 14  2806 non-null   object
 15  Bulk 15  2248 non-null   object
dtypes: int64(1), object(15)
memory usage: 391.2+ KB
None
****************************************************************************************************
data_gas_new:

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 3239 entries, 0 to 3238
Data columns (total 2 columns):
 #   Column  Non-Null Count  Dtype  
---  ------  --------------  -----  
 0   key     3239 non-null   int64  
 1   Газ 1   3239 non-null   float64
dtypes: float64(1), int64(1)
memory usage: 50.7 KB
None
****************************************************************************************************
data_temp_new:

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 18092 entries, 0 to 18091
Data columns (total 3 columns):
 #   Column        Non-Null Count  Dtype  
---  ------        --------------  -----  
 0   key           18092 non-null  int64  
 1   Время замера  18092 non-null  object 
 2   Температура   14665 non-null  float64
dtypes: float64(1), int64(1), object(1)
memory usage: 424.2+ KB
None
****************************************************************************************************
data_wire_new:

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 3081 entries, 0 to 3080
Data columns (total 10 columns):
 #   Column  Non-Null Count  Dtype  
---  ------  --------------  -----  
 0   key     3081 non-null   int64  
 1   Wire 1  3055 non-null   float64
 2   Wire 2  1079 non-null   float64
 3   Wire 3  63 non-null     float64
 4   Wire 4  14 non-null     float64
 5   Wire 5  1 non-null      float64
 6   Wire 6  73 non-null     float64
 7   Wire 7  11 non-null     float64
 8   Wire 8  19 non-null     float64
 9   Wire 9  29 non-null     float64
dtypes: float64(9), int64(1)
memory usage: 240.8 KB
None
****************************************************************************************************
data_wire_time_new:

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 3081 entries, 0 to 3080
Data columns (total 10 columns):
 #   Column  Non-Null Count  Dtype 
---  ------  --------------  ----- 
 0   key     3081 non-null   int64 
 1   Wire 1  3055 non-null   object
 2   Wire 2  1079 non-null   object
 3   Wire 3  63 non-null     object
 4   Wire 4  14 non-null     object
 5   Wire 5  1 non-null      object
 6   Wire 6  73 non-null     object
 7   Wire 7  11 non-null     object
 8   Wire 8  19 non-null     object
 9   Wire 9  29 non-null     object
dtypes: int64(1), object(9)
memory usage: 240.8+ KB
None
****************************************************************************************************


# Переведём время из формата строки в формат времени
df_arc['Начало нагрева дугой'] = pd.to_datetime(df_arc['Начало нагрева дугой'])
df_arc['Конец нагрева дугой'] = pd.to_datetime(df_arc['Конец нагрева дугой'])

df_arc.info()
df_arc.head()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 14876 entries, 0 to 14875
Data columns (total 5 columns):
 #   Column                Non-Null Count  Dtype         
---  ------                --------------  -----         
 0   key                   14876 non-null  int64         
 1   Начало нагрева дугой  14876 non-null  datetime64[ns]
 2   Конец нагрева дугой   14876 non-null  datetime64[ns]
 3   Активная мощность     14876 non-null  float64       
 4   Реактивная мощность   14876 non-null  float64       
dtypes: datetime64[ns](2), float64(2), int64(1)
memory usage: 581.2 KB


# Переведём время из формата строки в формат времени
bulk_columns = df_bulk_time.columns[1:]
for column in bulk_columns:
    df_bulk_time[column] = pd.to_datetime(df_bulk_time[column])

df_bulk_time.info()
df_bulk_time.head()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 3129 entries, 0 to 3128
Data columns (total 16 columns):
 #   Column   Non-Null Count  Dtype         
---  ------   --------------  -----         
 0   key      3129 non-null   int64         
 1   Bulk 1   252 non-null    datetime64[ns]
 2   Bulk 2   22 non-null     datetime64[ns]
 3   Bulk 3   1298 non-null   datetime64[ns]
 4   Bulk 4   1014 non-null   datetime64[ns]
 5   Bulk 5   77 non-null     datetime64[ns]
 6   Bulk 6   576 non-null    datetime64[ns]
 7   Bulk 7   25 non-null     datetime64[ns]
 8   Bulk 8   1 non-null      datetime64[ns]
 9   Bulk 9   19 non-null     datetime64[ns]
 10  Bulk 10  176 non-null    datetime64[ns]
 11  Bulk 11  177 non-null    datetime64[ns]
 12  Bulk 12  2450 non-null   datetime64[ns]
 13  Bulk 13  18 non-null     datetime64[ns]
 14  Bulk 14  2806 non-null   datetime64[ns]
 15  Bulk 15  2248 non-null   datetime64[ns]
dtypes: datetime64[ns](15), int64(1)
memory usage: 391.2 KB


# Переведём время из формата строки в формат времени
df_temp['Время замера'] = pd.to_datetime(df_temp['Время замера'])
df_temp['Время замера'] = pd.to_datetime(df_temp['Время замера'])

df_temp.info()
df_temp.head()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 18092 entries, 0 to 18091
Data columns (total 3 columns):
 #   Column        Non-Null Count  Dtype         
---  ------        --------------  -----         
 0   key           18092 non-null  int64         
 1   Время замера  18092 non-null  datetime64[ns]
 2   Температура   14665 non-null  float64       
dtypes: datetime64[ns](1), float64(1), int64(1)
memory usage: 424.2 KB


# Переведём время из формата строки в формат времени
wire_columns = df_wire_time.columns[1:]
for column in wire_columns:
    df_wire_time[column] = pd.to_datetime(df_wire_time[column])

df_wire_time.info()
df_wire_time.head()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 3081 entries, 0 to 3080
Data columns (total 10 columns):
 #   Column  Non-Null Count  Dtype         
---  ------  --------------  -----         
 0   key     3081 non-null   int64         
 1   Wire 1  3055 non-null   datetime64[ns]
 2   Wire 2  1079 non-null   datetime64[ns]
 3   Wire 3  63 non-null     datetime64[ns]
 4   Wire 4  14 non-null     datetime64[ns]
 5   Wire 5  1 non-null      datetime64[ns]
 6   Wire 6  73 non-null     datetime64[ns]
 7   Wire 7  11 non-null     datetime64[ns]
 8   Wire 8  19 non-null     datetime64[ns]
 9   Wire 9  29 non-null     datetime64[ns]
dtypes: datetime64[ns](9), int64(1)
memory usage: 240.8 KB


# Функция вывода уникальных значений столбцов
def get_unique_sorted(df):
    # Итерируемся по каждому столбцу в датафрейме
    for column in df.columns:
        # Получаем уникальные значения столбца и сортируем их
        unique_values = df[column].sort_values().unique()
# Выводим название столбца в кавычках и добавляем двоеточие
        print('Уникальные значения столбца', f"'{column}':")
# Выводим уникальные значения столбца
        print(unique_values)
# Выводим количество уникальных значений
        print(f'Количество уникальных значений: {len(unique_values)}')
# Делаем разделительную линию
        print('*'*100)


get_unique_sorted(df_arc)

Уникальные значения столбца 'key':
[   1    2    3 ... 3239 3240 3241]
Количество уникальных значений: 3214
****************************************************************************************************
Уникальные значения столбца 'Начало нагрева дугой':
['2019-05-03T11:02:14.000000000' '2019-05-03T11:07:28.000000000'
 '2019-05-03T11:11:44.000000000' ... '2019-09-06T17:06:48.000000000'
 '2019-09-06T17:21:58.000000000' '2019-09-06T17:24:54.000000000']
Количество уникальных значений: 14876
****************************************************************************************************
Уникальные значения столбца 'Конец нагрева дугой':
['2019-05-03T11:06:02.000000000' '2019-05-03T11:10:33.000000000'
 '2019-05-03T11:14:36.000000000' ... '2019-09-06T17:09:52.000000000'
 '2019-09-06T17:22:55.000000000' '2019-09-06T17:26:15.000000000']
Количество уникальных значений: 14876
****************************************************************************************************
Уникальные значения столбца 'Активная мощность':
[0.22312  0.223238 0.223895 ... 1.444904 1.458773 1.463773]
Количество уникальных значений: 13846
****************************************************************************************************
Уникальные значения столбца 'Реактивная мощность':
[-7.15479924e+02  1.53777000e-01  1.53921000e-01 ...  1.22306300e+00
  1.25862800e+00  1.27028400e+00]
Количество уникальных значений: 14707
****************************************************************************************************


# Сводная статистика для всех столбцов в `df_arc`
df_arc.describe(include='all', datetime_is_numeric=True)


# Среднее количество обработок для одной партии
print('Среднее количество обработок для одной партии: ', round(
    df_arc['key'].count() / df_arc['key'].nunique(), 2))

Среднее количество обработок для одной партии:  4.63


# Распределение партий по количеству обработок
df_count = df_arc.groupby('key')['key'].count().value_counts().reset_index()
df_count.columns = ['кол-во обработок', 'кол-во партий']
total_batches = df_count['кол-во партий'].sum()
df_count['% кол-ва партий'] = df_count['кол-во партий'] / total_batches * 100

plt.figure(figsize=(15, 5))
plt.barh(df_count['кол-во обработок'], df_count['кол-во партий'])
plt.xlabel('кол-во партий', fontsize=15, color='DarkSlateGray')
plt.ylabel('кол-во обработок', fontsize=15, color='DarkSlateGray')
plt.title('График количества обработок и партий',
          fontsize=20, color='DarkSlateGray')
plt.minorticks_on()
plt.grid(which='minor', linestyle=':')
plt.grid(True)
plt.show()
df_count


# Посмотрим как мощности распределены по датам
plt.figure(figsize=(15, 5))
plt.plot(df_arc['Начало нагрева дугой'],
         df_arc['Активная мощность'])
plt.title('Линейный график активной мощности по началу нагрева дугой',
          fontsize=15, color='DarkSlateGray')
plt.xlabel('Начало нагрева дугой',
           fontsize=12,
           color='DarkSlateGray')
plt.ylabel('Активная мощность',
           fontsize=12,
           color='DarkSlateGray')
plt.minorticks_on()
plt.grid(which='minor', linestyle=':')
plt.grid(True)
plt.show()


# Посмотрим на пустоту
px.line(df_arc, x='Конец нагрева дугой',
        y='Активная мощность', width=900, height=400).show()
px.line(df_arc, x='Начало нагрева дугой',
        y='Активная мощность', width=900, height=400).show()


# Записи с отрицательной реактивной мощностью
df_arc.query('`Реактивная мощность` < 0')


'''
Отфильтруем реактивную мощность меньше нуля и для остальных значений 
вычислим корреляцию между активной и реактивной мощностью
'''
df_arc.query('`Реактивная мощность` > 0')[
    ['Активная мощность', 'Реактивная мощность']].corr().round(2)


# Расчитаем реактивную мощность чтоб заменить отрицательное значение
df_arc.loc[df_arc['Реактивная мощность'] < 0,
           'Реактивная мощность'] = df_arc['Активная мощность'] * (df_arc.query('`Реактивная мощность` > 0')['Реактивная мощность'] /
                                                                   df_arc.query('`Реактивная мощность` > 0')['Активная мощность']).mean().round(6)


px.line(df_arc, x='Конец нагрева дугой',
        y='Реактивная мощность', width=900, height=400).show()
px.line(df_arc, x='Начало нагрева дугой',
        y='Реактивная мощность', width=900, height=400).show()


# Создаем графическую фигуру
plt.figure(figsize=(15, 8))
# Pазбиваем графическую фигуру на 2 графика
# Первый график
plt.subplot(2, 1, 1)
# Гистограмма для 'Активной мощности'
sns.histplot(data=df_arc['Активная мощность'],
             color='green', alpha=0.8, label='Активная мощность')
# Гистограмма для 'Реактивной мощности'
sns.histplot(data=df_arc['Реактивная мощность'],
             color='red', alpha=0.5, label='Реактивная мощность')

plt.minorticks_on()
plt.grid(which='minor', linestyle=':')
plt.grid(True)
plt.xlabel('')
plt.ylabel('Частота', fontsize=12, color='DarkSlateGray')
plt.legend()

# Второй график
plt.subplot(2, 1, 2)
# Боксплот для 'Активной мощности' и 'Реактивной мощности'
sns.boxplot(data=df_arc[['Активная мощность', 'Реактивная мощность']],
            orient='horizontal', palette=['green', 'red'])

plt.minorticks_on()
plt.grid(which='minor', linestyle=':')
plt.grid(True)
plt.xlabel('Мощность', fontsize=12, color='DarkSlateGray')
plt.ylabel('')
plt.suptitle('Распределение мощности (активной и реактивной)',
             fontsize=15, color='DarkSlateGray')
plt.show()


(df_arc['Конец нагрева дугой'] - df_arc['Начало нагрева дугой']).describe()

count                        14876
mean     0 days 00:02:51.670946490
std      0 days 00:01:38.186802680
min                0 days 00:00:11
25%                0 days 00:01:47
50%                0 days 00:02:27
75%                0 days 00:03:34
max                0 days 00:15:07
dtype: object


# Добавим столбец `Полная мощность`
df_arc['Полная мощность'] = np.sqrt(
    df_arc['Активная мощность']**2 + df_arc['Реактивная мощность']**2)


# Вычесляем продолжительность дуги
df_arc['Продолжительность дуги'] = (
    df_arc['Конец нагрева дугой'] - df_arc['Начало нагрева дугой']).astype('timedelta64[s]')
# Вычесляем количество потребляемой энергии
df_arc['Потребляемая энергия'] = df_arc['Полная мощность'] * \
    df_arc['Продолжительность дуги']


df_arc


agg_func_arc = {
    'key': 'count',
    'Начало нагрева дугой': 'first',
    'Конец нагрева дугой': 'last',
    'Активная мощность': 'sum',
    'Реактивная мощность': 'sum',
    'Полная мощность': 'sum',
    'Продолжительность дуги': 'sum',
    'Потребляемая энергия': 'sum'
}


df_arc_key = df_arc.groupby('key').agg(agg_func_arc).rename(columns={'key':'количество этапов обработки'})
df_arc_key


# Приведём таблицу к общепринятому виду
df_arc_key.columns = ['count_arc', 'start_arc', 'end_arc', 'active_power',
                      'reactive_power', 'apparent_power', 'arc_duration', 'energy_consumption']
df_arc_key


# Проверим столбец `key` на уникальность значений
df_bulk['key'].duplicated().sum()

0


# Столбец `key` установим в качестве индекса датафрейма
df_bulk = df_bulk.set_index('key')


# Сводная статистика для всех столбцов в `df_bulk`
df_bulk.describe()


# Посмотрим на уникальную запись с добавлением `Bulk 8`
df_bulk[df_bulk['Bulk 8'].notna()]


# Визуализируем распределения по колонкам `df_bulk`
plt.figure(figsize=(15, 10))
sns.boxplot(data=df_bulk, orient='horizontal')

plt.xlabel('Объем подачи сыпучих материалов',
           fontsize=12, color='DarkSlateGray')
plt.ylabel('Материал', fontsize=12, color='DarkSlateGray')
plt.title('Boxplot подачи сыпучих материалов',
          fontsize=15, color='DarkSlateGray')

plt.minorticks_on()
plt.grid(which='minor', linestyle=':')
plt.grid(True)


fig = px.box(df_bulk)
fig.update_layout(
    width=800, 
    height=500,
    xaxis=dict(title='Материал'),
    yaxis=dict(title='Объем подачи сыпучих материалов'),
    title='Boxplot подачи сыпучих материалов'
)
fig.update_xaxes(showgrid=True, gridwidth=1, gridcolor='lightgray')
fig.update_yaxes(showgrid=True, gridwidth=1, gridcolor='lightgray')

fig.show()


# Посмотрим уникальные значения по колонкам `df_bulk`
get_unique_sorted(df_bulk)

Уникальные значения столбца 'Bulk 1':
[ 10.  16.  19.  21.  22.  24.  26.  27.  28.  29.  30.  31.  32.  33.
  34.  36.  37.  38.  39.  41.  43.  44.  46.  47.  48.  50.  51.  52.
  54.  58.  60.  62.  63.  65.  67.  69.  73.  74.  75.  78.  83.  90.
  92.  94. 104. 118. 185.  nan]
Количество уникальных значений: 48
****************************************************************************************************
Уникальные значения столбца 'Bulk 2':
[228. 232. 233. 236. 242. 246. 247. 248. 249. 254. 257. 258. 270. 282.
 325.  nan]
Количество уникальных значений: 16
****************************************************************************************************
Уникальные значения столбца 'Bulk 3':
[  6.  17.  18.  19.  20.  21.  22.  23.  24.  25.  26.  27.  28.  29.
  30.  31.  32.  34.  36.  37.  38.  39.  40.  41.  42.  43.  44.  45.
  46.  47.  48.  49.  50.  51.  52.  53.  54.  55.  56.  57.  58.  59.
  60.  61.  62.  63.  64.  65.  66.  67.  68.  69.  70.  71.  72.  73.
  74.  75.  76.  77.  78.  79.  80.  81.  82.  83.  84.  85.  86.  87.
  88.  89.  90.  91.  92.  93.  94.  95.  96.  97.  98.  99. 100. 101.
 102. 103. 104. 105. 106. 107. 108. 109. 110. 111. 112. 113. 114. 115.
 116. 117. 118. 119. 120. 121. 122. 123. 124. 125. 126. 127. 128. 129.
 130. 131. 132. 133. 134. 135. 136. 137. 138. 139. 140. 141. 142. 143.
 144. 145. 146. 147. 148. 149. 150. 151. 152. 153. 154. 155. 156. 157.
 158. 159. 162. 163. 164. 165. 166. 167. 168. 169. 170. 171. 172. 173.
 174. 175. 176. 177. 178. 180. 181. 182. 183. 184. 185. 186. 187. 188.
 189. 190. 191. 193. 194. 196. 197. 198. 199. 200. 201. 202. 203. 204.
 206. 207. 208. 209. 210. 211. 212. 213. 214. 215. 220. 221. 222. 224.
 225. 226. 227. 230. 231. 232. 234. 235. 236. 237. 239. 240. 242. 243.
 244. 245. 246. 248. 249. 250. 251. 252. 255. 256. 258. 259. 260. 261.
 262. 263. 264. 265. 266. 267. 268. 269. 270. 272. 278. 280. 282. 284.
 285. 286. 289. 290. 291. 292. 294. 295. 297. 298. 300. 301. 302. 303.
 304. 305. 308. 310. 315. 316. 323. 324. 334. 335. 341. 342. 343. 349.
 352. 353. 354. 356. 358. 372. 379. 395. 405. 421. 445. 454.  nan]
Количество уникальных значений: 279
****************************************************************************************************
Уникальные значения столбца 'Bulk 4':
[ 12.  13.  16.  17.  19.  20.  21.  22.  23.  24.  25.  26.  27.  28.
  30.  31.  32.  33.  34.  36.  37.  38.  39.  40.  41.  42.  43.  44.
  45.  46.  47.  48.  49.  50.  51.  52.  53.  54.  55.  56.  57.  58.
  59.  60.  61.  62.  63.  64.  65.  66.  67.  68.  69.  70.  71.  72.
  73.  74.  75.  76.  77.  78.  79.  80.  81.  82.  83.  84.  85.  86.
  87.  88.  89.  90.  91.  92.  93.  94.  95.  96.  97.  98.  99. 100.
 101. 102. 103. 104. 105. 106. 107. 108. 109. 110. 111. 112. 113. 114.
 115. 116. 117. 118. 119. 120. 121. 122. 123. 124. 125. 126. 127. 128.
 129. 130. 131. 132. 133. 134. 135. 136. 137. 138. 139. 140. 141. 142.
 143. 144. 145. 146. 147. 148. 149. 150. 151. 152. 153. 154. 155. 156.
 157. 158. 159. 160. 161. 162. 163. 164. 165. 166. 167. 168. 170. 171.
 172. 174. 175. 176. 177. 179. 180. 181. 182. 184. 185. 186. 188. 189.
 190. 192. 193. 194. 195. 196. 198. 199. 200. 201. 202. 203. 205. 206.
 208. 212. 214. 218. 220. 221. 222. 223. 224. 228. 231. 233. 238. 239.
 241. 242. 243. 255. 256. 258. 265. 271. 276. 281.  nan]
Количество уникальных значений: 207
****************************************************************************************************
Уникальные значения столбца 'Bulk 5':
[ 11.  18.  19.  22.  23.  28.  31.  33.  34.  42.  43.  50.  54.  61.
  62.  70.  72.  73.  74.  77.  78.  79.  82.  83.  84.  86.  88.  90.
  94. 102. 103. 104. 112. 120. 121. 122. 124. 128. 132. 136. 142. 144.
 146. 148. 160. 180. 182. 184. 189. 197. 234. 242. 256. 293. 603.  nan]
Количество уникальных значений: 56
****************************************************************************************************
Уникальные значения столбца 'Bulk 6':
[ 17.  23.  24.  25.  26.  27.  28.  30.  32.  34.  36.  37.  38.  40.
  42.  43.  44.  45.  46.  47.  48.  49.  50.  51.  53.  54.  55.  56.
  57.  58.  59.  60.  61.  62.  63.  64.  65.  66.  67.  68.  69.  70.
  71.  72.  73.  74.  75.  76.  77.  78.  79.  80.  81.  82.  83.  84.
  85.  86.  87.  88.  89.  90.  91.  92.  93.  94.  95.  96.  97.  98.
  99. 100. 101. 102. 103. 104. 105. 106. 107. 108. 109. 110. 111. 112.
 114. 115. 116. 117. 118. 120. 121. 123. 124. 125. 126. 127. 128. 130.
 131. 132. 134. 135. 136. 137. 138. 139. 140. 141. 142. 143. 144. 145.
 146. 149. 150. 151. 152. 153. 155. 156. 157. 158. 159. 160. 162. 163.
 164. 165. 166. 167. 168. 170. 172. 173. 174. 175. 176. 177. 178. 179.
 180. 181. 182. 184. 185. 186. 187. 189. 190. 193. 194. 195. 196. 198.
 202. 203. 204. 206. 208. 211. 212. 215. 218. 219. 220. 224. 226. 230.
 231. 232. 234. 236. 240. 247. 250. 252. 254. 255. 261. 262. 264. 265.
 266. 271. 272. 278. 280. 283. 284. 285. 288. 294. 295. 297. 321. 325.
 326. 330. 350. 357. 359. 407. 414. 498. 503.  nan]
Количество уникальных значений: 206
****************************************************************************************************
Уникальные значения столбца 'Bulk 7':
[ 47.  50.  54.  75.  92. 108. 155. 178. 208. 250. 252. 296. 298. 306.
 340. 352. 392. 405. 406. 462. 506. 507. 553. 576. 772.  nan]
Количество уникальных значений: 26
****************************************************************************************************
Уникальные значения столбца 'Bulk 8':
[49. nan]
Количество уникальных значений: 2
****************************************************************************************************
Уникальные значения столбца 'Bulk 9':
[ 63.  65.  66.  68.  70.  71.  74. 108. 111. 147.  nan]
Количество уникальных значений: 11
****************************************************************************************************
Уникальные значения столбца 'Bulk 10':
[ 24.  26.  28.  30.  32.  34.  36.  38.  45.  46.  47.  49.  50.  52.
  54.  55.  56.  57.  58.  59.  61.  62.  63.  64.  65.  67.  68.  70.
  72.  73.  75.  76.  77.  79.  80.  81.  82.  83.  84.  85.  86.  87.
  88.  89.  90.  91.  92.  93.  95.  96.  97.  98.  99. 100. 101. 102.
 103. 104. 105. 106. 107. 108. 110. 111. 113. 114. 115. 117. 118. 122.
 124. 125. 126. 128. 147. 154. 159.  nan]
Количество уникальных значений: 78
****************************************************************************************************
Уникальные значения столбца 'Bulk 11':
[  8.  14.  16.  17.  18.  19.  20.  21.  22.  23.  24.  25.  26.  28.
  29.  31.  32.  34.  37.  38.  39.  40.  41.  42.  43.  46.  47.  48.
  49.  50.  51.  53.  54.  56.  58.  62.  64.  66.  67.  68.  69.  71.
  72.  74.  76.  81.  82.  83.  84.  87.  88.  89.  90.  91.  93.  94.
  95.  96.  97.  98. 100. 102. 103. 104. 106. 107. 114. 120. 124. 126.
 128. 131. 133. 136. 139. 140. 144. 146. 148. 151. 152. 154. 156. 158.
 162. 172. 174. 176. 178. 185. 188. 198. 200. 204. 205. 206. 211. 225.
 226. 242. 313.  nan]
Количество уникальных значений: 102
****************************************************************************************************
Уникальные значения столбца 'Bulk 12':
[  53.   54.   55.   56.   58.   88.  102.  103.  104.  105.  106.  107.
  108.  109.  110.  111.  115.  116.  123.  124.  125.  126.  127.  128.
  129.  135.  136.  137.  138.  139.  143.  145.  152.  153.  154.  155.
  156.  157.  158.  159.  160.  162.  165.  166.  167.  168.  170.  173.
  174.  175.  176.  177.  178.  180.  181.  182.  183.  184.  185.  186.
  187.  188.  189.  191.  192.  195.  201.  202.  203.  204.  205.  206.
  207.  208.  209.  210.  211.  212.  213.  214.  215.  216.  217.  219.
  220.  221.  222.  223.  224.  225.  226.  227.  228.  229.  230.  232.
  233.  234.  235.  236.  238.  239.  240.  242.  244.  247.  248.  250.
  252.  253.  254.  255.  256.  257.  258.  259.  260.  261.  262.  263.
  264.  265.  266.  267.  268.  270.  271.  272.  273.  274.  275.  276.
  277.  279.  280.  281.  282.  283.  284.  285.  286.  289.  290.  291.
  296.  301.  302.  303.  304.  305.  306.  307.  308.  309.  310.  311.
  312.  313.  314.  316.  319.  321.  322.  325.  329.  330.  331.  332.
  333.  334.  335.  337.  338.  340.  342.  343.  345.  346.  349.  351.
  352.  356.  358.  359.  360.  361.  362.  363.  364.  366.  368.  369.
  370.  371.  372.  373.  375.  377.  378.  379.  380.  381.  382.  383.
  385.  387.  388.  390.  391.  392.  393.  394.  395.  397.  399.  401.
  402.  404.  406.  407.  408.  409.  410.  411.  412.  413.  414.  415.
  416.  417.  418.  419.  420.  421.  422.  423.  424.  425.  428.  430.
  431.  432.  433.  434.  435.  436.  437.  439.  440.  441.  443.  444.
  445.  446.  447.  452.  453.  456.  457.  458.  459.  460.  461.  462.
  463.  464.  465.  466.  467.  468.  469.  470.  473.  474.  475.  478.
  479.  481.  483.  484.  496.  501.  507.  508.  509.  510.  511.  512.
  513.  514.  515.  516.  517.  518.  519.  520.  521.  522.  528.  542.
  550.  558.  561.  562.  563.  564.  567.  570.  572.  583.  586.  597.
  609.  616.  618.  620.  622.  624.  647.  666.  667.  668.  669.  671.
  685.  733.  774.  775.  833.  853. 1849.   nan]
Количество уникальных значений: 332
****************************************************************************************************
Уникальные значения столбца 'Bulk 13':
[151. 152. 153. 154. 155. 156. 157. 159. 202. 204. 206. 214. 278. 305.
  nan]
Количество уникальных значений: 15
****************************************************************************************************
Уникальные значения столбца 'Bulk 14':
[ 16.  18.  24.  29.  48.  52.  64.  68.  69.  72.  76.  78.  80.  82.
  83.  84.  85.  86.  88.  89.  90.  91.  92.  93.  94.  95.  96.  97.
  98.  99. 100. 101. 102. 103. 104. 105. 106. 107. 108. 109. 110. 111.
 112. 113. 114. 115. 116. 117. 118. 119. 120. 121. 122. 123. 124. 125.
 126. 127. 128. 129. 130. 131. 132. 133. 134. 135. 136. 137. 138. 139.
 140. 141. 142. 143. 144. 145. 146. 147. 148. 149. 150. 151. 152. 153.
 154. 155. 156. 157. 158. 159. 160. 161. 162. 163. 164. 165. 166. 167.
 168. 169. 170. 171. 173. 174. 176. 177. 178. 179. 180. 181. 182. 183.
 184. 185. 186. 187. 188. 189. 190. 191. 192. 193. 194. 195. 196. 197.
 198. 199. 200. 201. 202. 203. 204. 205. 206. 207. 208. 209. 210. 211.
 212. 213. 214. 215. 216. 217. 218. 219. 220. 221. 222. 223. 224. 225.
 226. 227. 228. 229. 230. 231. 232. 233. 234. 235. 236. 237. 238. 239.
 240. 241. 242. 243. 244. 245. 246. 247. 248. 249. 250. 251. 252. 253.
 254. 255. 256. 257. 258. 259. 260. 261. 262. 263. 264. 265. 266. 267.
 268. 269. 270. 271. 272. 274. 275. 276. 277. 278. 279. 280. 281. 282.
 283. 284. 286. 288. 289. 290. 291. 292. 294. 295. 296. 297. 298. 299.
 300. 301. 302. 303. 304. 305. 306. 307. 308. 309. 310. 311. 312. 313.
 315. 316. 317. 320. 321. 324. 325. 327. 330. 333. 336. 337. 342. 347.
 348. 349. 351. 352. 353. 354. 356. 366. 371. 375. 376. 379. 385. 389.
 390. 398. 400. 401. 402. 403. 405. 406. 407. 417. 428. 446. 450. 452.
 504. 552. 601. 636.  nan]
Количество уникальных значений: 285
****************************************************************************************************
Уникальные значения столбца 'Bulk 15':
[  1.  10.  44.  50.  51.  52.  53.  54.  55.  56.  58.  59.  63.  64.
  65.  69.  70.  72.  73.  74.  75.  85.  86.  87.  88.  99. 100. 101.
 102. 103. 104. 105. 106. 107. 108. 109. 110. 111. 112. 113. 114. 115.
 116. 117. 118. 123. 124. 125. 126. 127. 128. 134. 136. 137. 145. 147.
 148. 149. 150. 151. 152. 153. 154. 155. 156. 157. 158. 159. 160. 161.
 163. 164. 166. 167. 168. 169. 170. 171. 172. 174. 175. 176. 177. 178.
 179. 181. 183. 185. 186. 187. 188. 189. 190. 191. 192. 193. 194. 196.
 197. 198. 199. 200. 201. 202. 203. 204. 205. 206. 207. 208. 209. 210.
 211. 212. 213. 214. 215. 216. 217. 218. 219. 220. 221. 222. 223. 224.
 225. 226. 227. 228. 229. 230. 231. 232. 233. 234. 235. 236. 237. 238.
 239. 240. 246. 251. 258. 260. 261. 262. 264. 266. 290. 295. 320. 322.
 337. 405.  nan]
Количество уникальных значений: 157
****************************************************************************************************


# Заполним значения NaN в DataFrame нулями
df_bulk = df_bulk.fillna(0)
# Рассчитаем сумму каждой строки и создим новый столбец с названием "bulk_sum"
df_bulk['bulk_sum'] = df_bulk.sum(axis=1)
# Заменим пробелы в названиях столбцов на нижнее подчеркивание и приведем их к нижнему регистру
df_bulk.columns = [x.replace(' ','_').lower() for x in  df_bulk.columns]
df_bulk


# Проверим столбец `key` на уникальность значений
df_bulk_time['key'].duplicated().sum()

0


# Столбец `key` установим в качестве индекса датафрейма
df_bulk_time = df_bulk_time.set_index('key')


# Посмотрим уникальные значения по колонкам `df_bulk_time`
get_unique_sorted(df_bulk_time)

Уникальные значения столбца 'Bulk 1':
['2019-05-03T17:42:46.000000000' '2019-05-05T16:26:52.000000000'
 '2019-05-05T17:18:36.000000000' '2019-05-05T18:32:50.000000000'
 '2019-05-05T19:34:07.000000000' '2019-05-05T20:36:08.000000000'
 '2019-05-05T21:41:07.000000000' '2019-05-05T22:07:40.000000000'
 '2019-05-06T03:19:35.000000000' '2019-05-07T17:37:05.000000000'
 '2019-05-07T22:54:29.000000000' '2019-05-08T01:46:58.000000000'
 '2019-05-08T01:53:35.000000000' '2019-05-08T09:50:17.000000000'
 '2019-05-08T14:10:47.000000000' '2019-05-08T18:38:10.000000000'
 '2019-05-08T19:41:48.000000000' '2019-05-08T20:35:02.000000000'
 '2019-05-09T00:10:01.000000000' '2019-05-09T14:14:29.000000000'
 '2019-05-11T03:53:30.000000000' '2019-05-11T03:58:38.000000000'
 '2019-05-11T06:21:38.000000000' '2019-05-11T07:01:15.000000000'
 '2019-05-11T22:08:02.000000000' '2019-05-11T23:59:58.000000000'
 '2019-05-12T04:52:28.000000000' '2019-05-13T03:00:10.000000000'
 '2019-05-13T04:58:40.000000000' '2019-05-13T09:44:14.000000000'
 '2019-05-14T01:20:24.000000000' '2019-05-14T02:50:28.000000000'
 '2019-05-14T23:59:20.000000000' '2019-05-15T02:23:25.000000000'
 '2019-05-15T04:41:54.000000000' '2019-05-15T10:09:57.000000000'
 '2019-05-15T23:03:53.000000000' '2019-05-16T03:29:05.000000000'
 '2019-05-16T10:36:08.000000000' '2019-05-16T12:14:14.000000000'
 '2019-05-16T14:25:11.000000000' '2019-05-16T15:08:02.000000000'
 '2019-05-17T04:13:52.000000000' '2019-05-17T15:34:35.000000000'
 '2019-05-17T17:57:19.000000000' '2019-05-18T15:15:23.000000000'
 '2019-05-18T16:12:19.000000000' '2019-05-18T18:58:58.000000000'
 '2019-05-19T06:29:14.000000000' '2019-05-20T08:58:33.000000000'
 '2019-05-20T17:28:11.000000000' '2019-05-20T18:24:16.000000000'
 '2019-05-20T19:42:20.000000000' '2019-05-22T03:27:05.000000000'
 '2019-05-22T04:41:37.000000000' '2019-05-22T09:05:48.000000000'
 '2019-05-22T16:03:59.000000000' '2019-05-23T18:00:49.000000000'
 '2019-05-24T07:25:20.000000000' '2019-05-28T16:09:48.000000000'
 '2019-05-28T16:42:10.000000000' '2019-05-28T23:58:18.000000000'
 '2019-05-29T11:07:59.000000000' '2019-05-29T16:02:25.000000000'
 '2019-05-30T10:37:34.000000000' '2019-05-30T11:48:38.000000000'
 '2019-06-03T04:07:28.000000000' '2019-06-03T05:19:12.000000000'
 '2019-06-04T11:54:54.000000000' '2019-06-04T16:19:59.000000000'
 '2019-06-05T10:49:37.000000000' '2019-06-05T20:16:32.000000000'
 '2019-06-06T16:06:06.000000000' '2019-06-06T17:57:05.000000000'
 '2019-06-06T18:43:52.000000000' '2019-06-06T19:19:59.000000000'
 '2019-06-06T19:34:20.000000000' '2019-06-06T21:53:51.000000000'
 '2019-06-08T06:48:21.000000000' '2019-06-08T21:02:27.000000000'
 '2019-06-09T04:50:04.000000000' '2019-06-09T10:16:31.000000000'
 '2019-06-09T10:37:29.000000000' '2019-06-10T04:29:05.000000000'
 '2019-06-10T04:46:03.000000000' '2019-06-11T02:42:28.000000000'
 '2019-06-11T03:44:35.000000000' '2019-06-11T06:08:17.000000000'
 '2019-06-11T08:44:22.000000000' '2019-06-11T09:37:50.000000000'
 '2019-06-11T10:10:21.000000000' '2019-06-11T11:34:49.000000000'
 '2019-06-11T14:33:01.000000000' '2019-06-12T21:25:00.000000000'
 '2019-06-13T09:26:07.000000000' '2019-06-13T10:36:33.000000000'
 '2019-06-13T17:43:37.000000000' '2019-06-13T22:35:50.000000000'
 '2019-06-14T12:55:26.000000000' '2019-06-15T14:00:58.000000000'
 '2019-06-15T14:51:36.000000000' '2019-06-15T21:36:47.000000000'
 '2019-06-16T08:33:03.000000000' '2019-06-17T18:44:57.000000000'
 '2019-06-18T15:56:27.000000000' '2019-06-18T17:43:08.000000000'
 '2019-06-19T05:36:13.000000000' '2019-06-19T05:41:40.000000000'
 '2019-06-19T09:46:00.000000000' '2019-06-19T21:21:04.000000000'
 '2019-06-19T21:59:12.000000000' '2019-06-20T21:50:46.000000000'
 '2019-06-21T03:43:02.000000000' '2019-06-21T04:35:36.000000000'
 '2019-06-21T09:54:58.000000000' '2019-06-21T10:33:10.000000000'
 '2019-06-22T09:27:52.000000000' '2019-06-22T10:54:21.000000000'
 '2019-06-22T13:44:53.000000000' '2019-06-24T01:30:36.000000000'
 '2019-06-24T02:00:07.000000000' '2019-06-24T05:39:49.000000000'
 '2019-06-24T07:17:15.000000000' '2019-06-24T10:14:40.000000000'
 '2019-06-24T19:38:11.000000000' '2019-06-25T10:42:02.000000000'
 '2019-06-25T11:36:24.000000000' '2019-06-25T15:52:45.000000000'
 '2019-06-25T19:00:08.000000000' '2019-06-25T20:56:27.000000000'
 '2019-06-27T15:18:47.000000000' '2019-06-27T15:58:49.000000000'
 '2019-06-27T17:30:37.000000000' '2019-06-27T18:29:06.000000000'
 '2019-06-27T22:02:59.000000000' '2019-06-27T22:21:02.000000000'
 '2019-06-28T20:41:47.000000000' '2019-06-29T13:33:44.000000000'
 '2019-06-30T05:57:45.000000000' '2019-06-30T06:40:22.000000000'
 '2019-06-30T07:16:16.000000000' '2019-06-30T22:56:04.000000000'
 '2019-07-03T01:08:30.000000000' '2019-07-03T04:08:37.000000000'
 '2019-07-03T21:53:06.000000000' '2019-07-04T01:19:35.000000000'
 '2019-07-04T07:49:47.000000000' '2019-07-04T08:40:29.000000000'
 '2019-07-04T11:12:49.000000000' '2019-07-05T18:40:24.000000000'
 '2019-07-05T21:08:57.000000000' '2019-07-05T21:24:17.000000000'
 '2019-07-06T02:35:43.000000000' '2019-07-06T04:58:35.000000000'
 '2019-07-06T10:32:44.000000000' '2019-07-06T12:30:47.000000000'
 '2019-07-06T13:44:06.000000000' '2019-07-06T14:36:09.000000000'
 '2019-07-06T15:51:32.000000000' '2019-07-06T17:17:20.000000000'
 '2019-07-06T19:46:02.000000000' '2019-07-07T22:56:51.000000000'
 '2019-07-09T04:01:44.000000000' '2019-07-09T21:25:01.000000000'
 '2019-07-09T22:56:41.000000000' '2019-07-10T02:37:50.000000000'
 '2019-07-10T16:29:50.000000000' '2019-07-11T15:05:01.000000000'
 '2019-07-11T17:09:01.000000000' '2019-07-12T18:07:10.000000000'
 '2019-07-12T18:53:16.000000000' '2019-07-12T19:29:31.000000000'
 '2019-07-13T02:20:59.000000000' '2019-07-13T03:04:10.000000000'
 '2019-07-23T03:06:03.000000000' '2019-07-23T10:17:34.000000000'
 '2019-07-23T11:23:22.000000000' '2019-07-26T22:47:29.000000000'
 '2019-07-27T04:17:21.000000000' '2019-07-27T05:43:35.000000000'
 '2019-07-27T06:23:07.000000000' '2019-07-27T08:54:09.000000000'
 '2019-07-27T16:34:15.000000000' '2019-07-27T18:12:32.000000000'
 '2019-07-27T19:11:22.000000000' '2019-07-28T14:11:24.000000000'
 '2019-07-29T16:07:12.000000000' '2019-07-30T19:39:56.000000000'
 '2019-07-30T21:55:33.000000000' '2019-08-01T06:06:34.000000000'
 '2019-08-02T23:33:57.000000000' '2019-08-03T04:25:56.000000000'
 '2019-08-04T00:29:49.000000000' '2019-08-05T03:45:25.000000000'
 '2019-08-05T10:10:45.000000000' '2019-08-08T23:30:42.000000000'
 '2019-08-08T23:54:19.000000000' '2019-08-09T15:49:15.000000000'
 '2019-08-09T17:36:23.000000000' '2019-08-11T02:14:16.000000000'
 '2019-08-12T15:51:11.000000000' '2019-08-12T19:51:09.000000000'
 '2019-08-12T23:16:11.000000000' '2019-08-13T01:07:34.000000000'
 '2019-08-13T01:36:24.000000000' '2019-08-13T04:37:28.000000000'
 '2019-08-13T07:29:19.000000000' '2019-08-13T09:18:00.000000000'
 '2019-08-13T10:02:17.000000000' '2019-08-13T12:02:20.000000000'
 '2019-08-14T03:53:23.000000000' '2019-08-14T07:06:34.000000000'
 '2019-08-14T09:12:31.000000000' '2019-08-14T15:38:48.000000000'
 '2019-08-14T16:26:27.000000000' '2019-08-14T18:07:06.000000000'
 '2019-08-19T02:38:32.000000000' '2019-08-20T11:19:12.000000000'
 '2019-08-20T16:14:40.000000000' '2019-08-21T15:21:58.000000000'
 '2019-08-23T19:10:47.000000000' '2019-08-24T17:19:07.000000000'
 '2019-08-25T01:05:20.000000000' '2019-08-25T20:57:58.000000000'
 '2019-08-26T19:53:39.000000000' '2019-08-26T20:56:06.000000000'
 '2019-08-26T21:27:45.000000000' '2019-08-27T00:25:48.000000000'
 '2019-08-27T11:24:10.000000000' '2019-08-27T12:02:50.000000000'
 '2019-08-27T16:12:26.000000000' '2019-08-28T08:33:28.000000000'
 '2019-08-28T09:34:30.000000000' '2019-08-29T15:19:43.000000000'
 '2019-08-30T12:07:47.000000000' '2019-08-31T18:55:07.000000000'
 '2019-08-31T19:44:46.000000000' '2019-09-01T17:59:26.000000000'
 '2019-09-01T18:45:39.000000000' '2019-09-01T19:08:00.000000000'
 '2019-09-02T16:40:49.000000000' '2019-09-02T22:44:37.000000000'
 '2019-09-03T09:24:41.000000000' '2019-09-04T04:48:19.000000000'
 '2019-09-04T16:01:08.000000000' '2019-09-04T19:33:27.000000000'
 '2019-09-04T20:39:01.000000000' '2019-09-04T20:51:19.000000000'
 '2019-09-05T01:56:45.000000000' '2019-09-05T02:37:08.000000000'
 '2019-09-05T06:10:59.000000000' '2019-09-05T09:11:32.000000000'
                           'NaT']
Количество уникальных значений: 253
****************************************************************************************************
Уникальные значения столбца 'Bulk 2':
['2019-05-07T15:39:35.000000000' '2019-05-07T16:16:34.000000000'
 '2019-05-07T19:13:07.000000000' '2019-05-07T22:50:29.000000000'
 '2019-05-08T01:46:58.000000000' '2019-05-08T01:53:35.000000000'
 '2019-07-27T04:17:21.000000000' '2019-07-27T05:35:32.000000000'
 '2019-07-27T06:23:07.000000000' '2019-07-27T08:54:09.000000000'
 '2019-07-27T16:24:45.000000000' '2019-07-27T18:12:32.000000000'
 '2019-07-27T19:11:22.000000000' '2019-08-12T21:42:20.000000000'
 '2019-08-12T21:48:11.000000000' '2019-08-13T01:01:50.000000000'
 '2019-08-13T02:46:14.000000000' '2019-08-13T04:37:28.000000000'
 '2019-08-13T07:26:20.000000000' '2019-08-13T09:14:04.000000000'
 '2019-08-13T09:55:35.000000000' '2019-08-13T11:47:39.000000000'
                           'NaT']
Количество уникальных значений: 23
****************************************************************************************************
Уникальные значения столбца 'Bulk 3':
['2019-05-03T20:40:25.000000000' '2019-05-04T04:25:29.000000000'
 '2019-05-04T05:15:12.000000000' ... '2019-09-06T11:54:15.000000000'
 '2019-09-06T12:26:52.000000000'                           'NaT']
Количество уникальных значений: 1299
****************************************************************************************************
Уникальные значения столбца 'Bulk 4':
['2019-05-03T11:28:48.000000000' '2019-05-03T11:36:50.000000000'
 '2019-05-03T12:32:39.000000000' ... '2019-09-05T03:12:35.000000000'
 '2019-09-05T03:35:21.000000000'                           'NaT']
Количество уникальных значений: 1015
****************************************************************************************************
Уникальные значения столбца 'Bulk 5':
['2019-05-07T15:19:17.000000000' '2019-05-07T16:46:56.000000000'
 '2019-05-07T19:04:09.000000000' '2019-05-07T22:32:04.000000000'
 '2019-05-08T01:25:11.000000000' '2019-05-08T01:53:35.000000000'
 '2019-05-09T16:32:49.000000000' '2019-05-12T09:17:31.000000000'
 '2019-05-12T09:50:33.000000000' '2019-05-12T10:54:36.000000000'
 '2019-05-12T11:40:43.000000000' '2019-05-12T12:25:57.000000000'
 '2019-05-12T13:06:07.000000000' '2019-05-12T15:10:42.000000000'
 '2019-05-17T03:23:34.000000000' '2019-05-22T21:39:11.000000000'
 '2019-06-09T06:13:50.000000000' '2019-06-16T00:51:58.000000000'
 '2019-06-19T22:37:16.000000000' '2019-06-19T23:32:44.000000000'
 '2019-07-04T13:44:47.000000000' '2019-07-04T14:00:31.000000000'
 '2019-07-04T15:17:45.000000000' '2019-07-04T15:55:32.000000000'
 '2019-07-04T16:48:21.000000000' '2019-07-04T18:15:19.000000000'
 '2019-07-04T19:44:56.000000000' '2019-07-04T20:19:26.000000000'
 '2019-07-04T20:58:58.000000000' '2019-07-07T18:35:15.000000000'
 '2019-07-07T19:21:56.000000000' '2019-07-07T20:44:45.000000000'
 '2019-07-07T22:15:12.000000000' '2019-07-09T14:36:47.000000000'
 '2019-07-10T03:27:50.000000000' '2019-07-10T04:28:19.000000000'
 '2019-07-25T16:00:24.000000000' '2019-07-25T16:58:47.000000000'
 '2019-07-25T17:59:41.000000000' '2019-07-25T18:46:56.000000000'
 '2019-07-25T19:36:40.000000000' '2019-07-25T20:44:18.000000000'
 '2019-07-25T21:39:04.000000000' '2019-07-27T04:10:08.000000000'
 '2019-07-27T05:21:58.000000000' '2019-07-27T06:45:47.000000000'
 '2019-07-27T09:53:31.000000000' '2019-07-27T15:29:14.000000000'
 '2019-07-27T18:12:32.000000000' '2019-07-27T19:11:22.000000000'
 '2019-08-02T19:22:26.000000000' '2019-08-02T20:29:54.000000000'
 '2019-08-02T21:29:51.000000000' '2019-08-12T21:20:02.000000000'
 '2019-08-12T22:28:59.000000000' '2019-08-12T23:49:50.000000000'
 '2019-08-13T01:45:01.000000000' '2019-08-13T04:23:23.000000000'
 '2019-08-13T08:05:11.000000000' '2019-08-13T09:51:02.000000000'
 '2019-08-13T11:04:10.000000000' '2019-08-17T04:59:15.000000000'
 '2019-08-17T05:36:00.000000000' '2019-08-17T06:00:38.000000000'
 '2019-08-24T15:41:14.000000000' '2019-09-01T08:31:55.000000000'
 '2019-09-01T09:20:27.000000000' '2019-09-01T09:37:36.000000000'
 '2019-09-02T06:33:55.000000000' '2019-09-02T07:28:42.000000000'
 '2019-09-02T08:28:58.000000000' '2019-09-02T14:12:23.000000000'
 '2019-09-02T15:08:19.000000000' '2019-09-02T16:19:39.000000000'
 '2019-09-02T16:58:07.000000000' '2019-09-02T17:34:47.000000000'
 '2019-09-02T18:16:52.000000000'                           'NaT']
Количество уникальных значений: 78
****************************************************************************************************
Уникальные значения столбца 'Bulk 6':
['2019-05-03T19:09:15.000000000' '2019-05-03T21:36:10.000000000'
 '2019-05-03T23:40:36.000000000' '2019-05-04T01:14:08.000000000'
 '2019-05-04T07:30:19.000000000' '2019-05-05T02:28:41.000000000'
 '2019-05-05T03:51:16.000000000' '2019-05-05T07:16:45.000000000'
 '2019-05-05T07:34:21.000000000' '2019-05-05T08:08:36.000000000'
 '2019-05-05T09:53:40.000000000' '2019-05-05T11:21:49.000000000'
 '2019-05-05T16:45:30.000000000' '2019-05-05T17:12:58.000000000'
 '2019-05-05T18:28:05.000000000' '2019-05-05T19:34:07.000000000'
 '2019-05-05T20:36:08.000000000' '2019-05-05T21:08:45.000000000'
 '2019-05-05T22:15:29.000000000' '2019-05-05T23:46:08.000000000'
 '2019-05-06T01:12:46.000000000' '2019-05-06T09:04:53.000000000'
 '2019-05-06T09:33:02.000000000' '2019-05-06T16:15:42.000000000'
 '2019-05-06T17:43:26.000000000' '2019-05-06T18:17:34.000000000'
 '2019-05-06T18:47:13.000000000' '2019-05-06T20:43:05.000000000'
 '2019-05-06T21:23:02.000000000' '2019-05-06T21:56:33.000000000'
 '2019-05-06T22:34:09.000000000' '2019-05-06T23:28:31.000000000'
 '2019-05-06T23:57:16.000000000' '2019-05-07T09:53:19.000000000'
 '2019-05-08T15:46:06.000000000' '2019-05-08T17:12:01.000000000'
 '2019-05-09T09:03:16.000000000' '2019-05-09T10:21:02.000000000'
 '2019-05-09T12:26:15.000000000' '2019-05-09T14:35:20.000000000'
 '2019-05-09T17:29:56.000000000' '2019-05-09T21:08:08.000000000'
 '2019-05-09T22:00:42.000000000' '2019-05-09T23:43:07.000000000'
 '2019-05-10T00:48:12.000000000' '2019-05-10T03:33:42.000000000'
 '2019-05-10T04:28:55.000000000' '2019-05-10T05:01:31.000000000'
 '2019-05-10T05:41:51.000000000' '2019-05-10T06:33:28.000000000'
 '2019-05-10T07:08:59.000000000' '2019-05-10T07:42:42.000000000'
 '2019-05-10T14:54:38.000000000' '2019-05-10T15:40:53.000000000'
 '2019-05-10T18:36:30.000000000' '2019-05-10T20:00:57.000000000'
 '2019-05-10T21:05:18.000000000' '2019-05-11T04:47:58.000000000'
 '2019-05-12T07:35:31.000000000' '2019-05-12T17:18:24.000000000'
 '2019-05-12T23:19:20.000000000' '2019-05-13T13:38:42.000000000'
 '2019-05-13T14:46:59.000000000' '2019-05-14T03:59:25.000000000'
 '2019-05-14T08:12:04.000000000' '2019-05-14T08:51:49.000000000'
 '2019-05-14T10:13:37.000000000' '2019-05-14T18:02:52.000000000'
 '2019-05-15T09:16:29.000000000' '2019-05-15T10:22:01.000000000'
 '2019-05-15T11:46:00.000000000' '2019-05-15T17:06:40.000000000'
 '2019-05-15T18:59:06.000000000' '2019-05-15T20:16:11.000000000'
 '2019-05-16T11:59:36.000000000' '2019-05-16T13:17:56.000000000'
 '2019-05-16T19:45:42.000000000' '2019-05-16T20:15:55.000000000'
 '2019-05-16T20:37:57.000000000' '2019-05-16T21:46:44.000000000'
 '2019-05-17T04:09:53.000000000' '2019-05-17T05:29:07.000000000'
 '2019-05-17T06:45:31.000000000' '2019-05-17T11:14:08.000000000'
 '2019-05-18T23:22:24.000000000' '2019-05-19T00:02:46.000000000'
 '2019-05-21T16:50:43.000000000' '2019-05-21T18:04:09.000000000'
 '2019-05-22T16:03:59.000000000' '2019-05-22T17:46:40.000000000'
 '2019-05-23T03:25:56.000000000' '2019-05-23T04:39:16.000000000'
 '2019-05-23T09:06:35.000000000' '2019-05-23T11:04:16.000000000'
 '2019-05-23T13:17:58.000000000' '2019-05-23T14:29:51.000000000'
 '2019-05-24T03:42:25.000000000' '2019-05-24T07:30:55.000000000'
 '2019-05-24T08:36:38.000000000' '2019-05-24T09:39:47.000000000'
 '2019-05-24T10:10:29.000000000' '2019-05-24T10:48:47.000000000'
 '2019-05-24T12:37:15.000000000' '2019-05-24T13:18:31.000000000'
 '2019-05-24T14:02:20.000000000' '2019-05-25T04:18:28.000000000'
 '2019-05-25T05:49:41.000000000' '2019-05-25T06:34:07.000000000'
 '2019-05-25T13:59:07.000000000' '2019-05-26T08:53:54.000000000'
 '2019-05-26T09:50:04.000000000' '2019-05-27T07:06:16.000000000'
 '2019-05-27T07:59:46.000000000' '2019-05-27T08:45:17.000000000'
 '2019-05-27T13:24:55.000000000' '2019-05-27T14:08:25.000000000'
 '2019-05-27T14:31:36.000000000' '2019-05-28T17:37:46.000000000'
 '2019-05-29T00:41:42.000000000' '2019-05-30T09:08:56.000000000'
 '2019-05-30T10:25:59.000000000' '2019-05-30T11:43:40.000000000'
 '2019-05-30T12:44:16.000000000' '2019-05-30T23:50:21.000000000'
 '2019-05-31T01:07:11.000000000' '2019-05-31T02:02:19.000000000'
 '2019-05-31T03:03:17.000000000' '2019-05-31T04:17:23.000000000'
 '2019-06-03T06:13:41.000000000' '2019-06-03T10:15:34.000000000'
 '2019-06-03T19:20:11.000000000' '2019-06-04T02:59:11.000000000'
 '2019-06-04T06:03:31.000000000' '2019-06-04T07:14:53.000000000'
 '2019-06-04T08:32:00.000000000' '2019-06-04T10:29:14.000000000'
 '2019-06-04T10:59:37.000000000' '2019-06-05T12:22:39.000000000'
 '2019-06-05T21:06:26.000000000' '2019-06-05T23:30:58.000000000'
 '2019-06-06T00:32:17.000000000' '2019-06-06T01:21:24.000000000'
 '2019-06-06T03:09:43.000000000' '2019-06-06T11:51:00.000000000'
 '2019-06-07T02:24:34.000000000' '2019-06-07T08:47:00.000000000'
 '2019-06-07T13:40:45.000000000' '2019-06-07T14:58:56.000000000'
 '2019-06-07T16:14:36.000000000' '2019-06-07T16:37:52.000000000'
 '2019-06-07T18:17:49.000000000' '2019-06-07T19:18:22.000000000'
 '2019-06-07T21:59:11.000000000' '2019-06-07T23:55:43.000000000'
 '2019-06-08T00:43:43.000000000' '2019-06-08T01:33:26.000000000'
 '2019-06-08T12:40:52.000000000' '2019-06-08T20:24:04.000000000'
 '2019-06-09T04:21:00.000000000' '2019-06-09T05:57:53.000000000'
 '2019-06-09T08:41:04.000000000' '2019-06-09T09:02:23.000000000'
 '2019-06-09T10:16:31.000000000' '2019-06-09T10:42:54.000000000'
 '2019-06-11T07:07:17.000000000' '2019-06-11T08:32:37.000000000'
 '2019-06-11T08:57:20.000000000' '2019-06-11T09:33:20.000000000'
 '2019-06-11T10:25:40.000000000' '2019-06-11T11:42:54.000000000'
 '2019-06-11T13:02:20.000000000' '2019-06-11T15:00:48.000000000'
 '2019-06-11T16:06:08.000000000' '2019-06-11T16:16:49.000000000'
 '2019-06-11T18:53:24.000000000' '2019-06-11T19:30:24.000000000'
 '2019-06-11T20:41:43.000000000' '2019-06-11T21:53:17.000000000'
 '2019-06-11T22:38:01.000000000' '2019-06-11T23:01:45.000000000'
 '2019-06-11T23:47:56.000000000' '2019-06-12T00:10:33.000000000'
 '2019-06-12T05:43:39.000000000' '2019-06-12T07:04:01.000000000'
 '2019-06-12T16:03:03.000000000' '2019-06-12T17:24:27.000000000'
 '2019-06-12T17:48:44.000000000' '2019-06-12T19:18:54.000000000'
 '2019-06-12T19:34:53.000000000' '2019-06-12T20:58:31.000000000'
 '2019-06-14T10:41:51.000000000' '2019-06-14T11:30:08.000000000'
 '2019-06-14T12:08:21.000000000' '2019-06-14T12:37:53.000000000'
 '2019-06-14T13:30:53.000000000' '2019-06-14T14:01:23.000000000'
 '2019-06-14T15:28:13.000000000' '2019-06-14T20:24:48.000000000'
 '2019-06-15T12:54:38.000000000' '2019-06-17T02:23:02.000000000'
 '2019-06-17T03:07:37.000000000' '2019-06-17T04:07:05.000000000'
 '2019-06-17T04:20:40.000000000' '2019-06-17T05:36:23.000000000'
 '2019-06-17T06:50:27.000000000' '2019-06-17T07:19:09.000000000'
 '2019-06-17T08:49:20.000000000' '2019-06-17T10:05:58.000000000'
 '2019-06-17T10:26:50.000000000' '2019-06-17T11:19:04.000000000'
 '2019-06-18T12:12:46.000000000' '2019-06-18T20:06:39.000000000'
 '2019-06-20T11:17:48.000000000' '2019-06-21T10:33:10.000000000'
 '2019-06-22T19:42:32.000000000' '2019-06-23T02:12:27.000000000'
 '2019-06-23T03:24:26.000000000' '2019-06-23T10:13:01.000000000'
 '2019-06-23T22:45:40.000000000' '2019-06-23T22:54:06.000000000'
 '2019-06-24T00:35:20.000000000' '2019-06-24T01:24:22.000000000'
 '2019-06-24T02:04:09.000000000' '2019-06-24T03:26:53.000000000'
 '2019-06-24T04:13:08.000000000' '2019-06-24T05:37:36.000000000'
 '2019-06-24T12:51:21.000000000' '2019-06-25T01:58:37.000000000'
 '2019-06-25T13:49:21.000000000' '2019-06-25T15:24:29.000000000'
 '2019-06-25T16:05:49.000000000' '2019-06-25T17:37:56.000000000'
 '2019-06-25T18:13:24.000000000' '2019-06-25T18:49:04.000000000'
 '2019-06-26T07:46:42.000000000' '2019-06-28T21:50:34.000000000'
 '2019-06-29T00:14:17.000000000' '2019-06-29T01:13:14.000000000'
 '2019-06-29T01:56:12.000000000' '2019-06-29T02:34:20.000000000'
 '2019-06-29T03:05:28.000000000' '2019-06-29T03:43:16.000000000'
 '2019-06-29T04:41:33.000000000' '2019-06-29T05:33:04.000000000'
 '2019-06-29T08:30:56.000000000' '2019-06-30T06:47:04.000000000'
 '2019-07-01T10:12:14.000000000' '2019-07-02T00:04:02.000000000'
 '2019-07-02T00:45:12.000000000' '2019-07-02T01:20:12.000000000'
 '2019-07-02T09:58:54.000000000' '2019-07-02T11:32:19.000000000'
 '2019-07-02T12:37:09.000000000' '2019-07-02T13:08:19.000000000'
 '2019-07-02T15:11:54.000000000' '2019-07-02T15:40:45.000000000'
 '2019-07-02T16:12:58.000000000' '2019-07-02T17:24:21.000000000'
 '2019-07-02T18:19:31.000000000' '2019-07-02T18:29:36.000000000'
 '2019-07-02T19:50:12.000000000' '2019-07-02T20:05:38.000000000'
 '2019-07-02T21:04:43.000000000' '2019-07-02T21:21:53.000000000'
 '2019-07-02T22:14:21.000000000' '2019-07-02T22:44:02.000000000'
 '2019-07-04T11:46:51.000000000' '2019-07-04T13:25:12.000000000'
 '2019-07-04T14:00:31.000000000' '2019-07-04T15:13:33.000000000'
 '2019-07-04T15:48:59.000000000' '2019-07-04T16:43:53.000000000'
 '2019-07-04T17:36:09.000000000' '2019-07-04T19:12:16.000000000'
 '2019-07-04T19:44:56.000000000' '2019-07-04T20:14:06.000000000'
 '2019-07-04T20:55:52.000000000' '2019-07-05T14:39:01.000000000'
 '2019-07-05T15:15:52.000000000' '2019-07-05T16:23:29.000000000'
 '2019-07-05T16:46:22.000000000' '2019-07-05T21:44:54.000000000'
 '2019-07-05T22:20:06.000000000' '2019-07-07T18:29:55.000000000'
 '2019-07-07T19:40:36.000000000' '2019-07-07T20:55:36.000000000'
 '2019-07-07T23:05:10.000000000' '2019-07-09T02:15:18.000000000'
 '2019-07-09T04:54:37.000000000' '2019-07-10T08:08:55.000000000'
 '2019-07-10T09:25:54.000000000' '2019-07-10T10:58:30.000000000'
 '2019-07-11T13:48:19.000000000' '2019-07-12T03:28:59.000000000'
 '2019-07-12T04:11:51.000000000' '2019-07-12T10:13:21.000000000'
 '2019-07-12T13:48:37.000000000' '2019-07-12T22:54:35.000000000'
 '2019-07-13T05:38:33.000000000' '2019-07-13T07:00:40.000000000'
 '2019-07-18T09:21:22.000000000' '2019-07-19T07:38:09.000000000'
 '2019-07-19T09:11:21.000000000' '2019-07-19T10:18:05.000000000'
 '2019-07-19T13:16:52.000000000' '2019-07-19T14:23:48.000000000'
 '2019-07-19T15:31:37.000000000' '2019-07-19T16:51:28.000000000'
 '2019-07-19T18:00:24.000000000' '2019-07-19T19:00:09.000000000'
 '2019-07-19T19:34:07.000000000' '2019-07-19T20:02:03.000000000'
 '2019-07-19T21:03:34.000000000' '2019-07-19T22:20:34.000000000'
 '2019-07-19T23:19:45.000000000' '2019-07-20T01:48:13.000000000'
 '2019-07-20T04:22:08.000000000' '2019-07-20T06:16:56.000000000'
 '2019-07-20T07:45:21.000000000' '2019-07-20T14:26:09.000000000'
 '2019-07-20T15:32:47.000000000' '2019-07-20T20:12:23.000000000'
 '2019-07-21T00:25:45.000000000' '2019-07-21T14:39:15.000000000'
 '2019-07-21T23:56:15.000000000' '2019-07-22T05:15:42.000000000'
 '2019-07-22T06:47:50.000000000' '2019-07-22T07:23:48.000000000'
 '2019-07-22T07:56:47.000000000' '2019-07-22T17:27:34.000000000'
 '2019-07-23T05:15:08.000000000' '2019-07-23T15:43:15.000000000'
 '2019-07-24T07:08:45.000000000' '2019-07-24T07:57:08.000000000'
 '2019-07-24T09:00:12.000000000' '2019-07-24T09:20:43.000000000'
 '2019-07-24T10:19:10.000000000' '2019-07-24T10:34:50.000000000'
 '2019-07-24T11:17:22.000000000' '2019-07-24T12:03:28.000000000'
 '2019-07-24T12:51:24.000000000' '2019-07-24T13:57:14.000000000'
 '2019-07-24T15:43:34.000000000' '2019-07-24T17:07:06.000000000'
 '2019-07-24T17:55:25.000000000' '2019-07-24T18:47:47.000000000'
 '2019-07-24T20:22:55.000000000' '2019-07-24T20:51:46.000000000'
 '2019-07-24T21:37:54.000000000' '2019-07-24T22:03:04.000000000'
 '2019-07-24T22:56:18.000000000' '2019-07-24T23:33:50.000000000'
 '2019-07-24T23:47:11.000000000' '2019-07-25T01:01:21.000000000'
 '2019-07-25T15:53:11.000000000' '2019-07-26T09:56:39.000000000'
 '2019-07-26T19:27:20.000000000' '2019-07-26T20:35:54.000000000'
 '2019-07-26T23:46:09.000000000' '2019-07-27T00:04:49.000000000'
 '2019-07-27T00:34:20.000000000' '2019-07-27T01:04:11.000000000'
 '2019-07-27T02:13:58.000000000' '2019-07-27T23:57:05.000000000'
 '2019-07-28T01:40:36.000000000' '2019-07-28T02:30:28.000000000'
 '2019-07-28T03:17:47.000000000' '2019-07-28T04:31:51.000000000'
 '2019-07-28T05:41:19.000000000' '2019-07-28T06:11:49.000000000'
 '2019-07-28T22:45:55.000000000' '2019-07-29T02:15:35.000000000'
 '2019-07-29T02:41:51.000000000' '2019-07-29T03:44:00.000000000'
 '2019-07-29T04:44:11.000000000' '2019-07-29T06:21:27.000000000'
 '2019-07-30T05:23:52.000000000' '2019-07-30T14:02:12.000000000'
 '2019-07-30T22:41:55.000000000' '2019-07-30T23:58:22.000000000'
 '2019-07-31T00:16:44.000000000' '2019-07-31T01:13:17.000000000'
 '2019-07-31T02:17:31.000000000' '2019-07-31T03:52:07.000000000'
 '2019-07-31T04:03:30.000000000' '2019-07-31T05:47:04.000000000'
 '2019-07-31T13:38:49.000000000' '2019-08-01T01:53:31.000000000'
 '2019-08-01T02:33:22.000000000' '2019-08-01T03:51:46.000000000'
 '2019-08-01T05:35:27.000000000' '2019-08-01T06:06:34.000000000'
 '2019-08-01T06:55:32.000000000' '2019-08-01T07:47:18.000000000'
 '2019-08-02T03:50:22.000000000' '2019-08-02T04:32:32.000000000'
 '2019-08-02T05:04:32.000000000' '2019-08-02T19:41:40.000000000'
 '2019-08-03T02:51:16.000000000' '2019-08-03T04:47:06.000000000'
 '2019-08-03T05:40:35.000000000' '2019-08-04T15:11:10.000000000'
 '2019-08-04T18:43:15.000000000' '2019-08-05T06:58:07.000000000'
 '2019-08-05T07:40:00.000000000' '2019-08-05T08:50:29.000000000'
 '2019-08-05T10:03:23.000000000' '2019-08-05T10:24:28.000000000'
 '2019-08-05T11:01:20.000000000' '2019-08-05T11:31:20.000000000'
 '2019-08-05T13:49:09.000000000' '2019-08-06T04:10:17.000000000'
 '2019-08-06T05:19:43.000000000' '2019-08-06T07:07:25.000000000'
 '2019-08-06T07:28:26.000000000' '2019-08-06T08:37:44.000000000'
 '2019-08-06T09:00:32.000000000' '2019-08-06T10:15:15.000000000'
 '2019-08-06T10:58:23.000000000' '2019-08-06T11:34:02.000000000'
 '2019-08-06T12:10:13.000000000' '2019-08-06T12:31:18.000000000'
 '2019-08-06T13:35:35.000000000' '2019-08-06T14:00:06.000000000'
 '2019-08-06T15:30:27.000000000' '2019-08-07T06:31:05.000000000'
 '2019-08-07T16:00:31.000000000' '2019-08-07T16:16:00.000000000'
 '2019-08-07T17:04:21.000000000' '2019-08-07T17:40:33.000000000'
 '2019-08-07T18:14:48.000000000' '2019-08-07T18:48:39.000000000'
 '2019-08-07T19:14:08.000000000' '2019-08-07T21:00:46.000000000'
 '2019-08-07T21:18:41.000000000' '2019-08-07T22:03:53.000000000'
 '2019-08-07T22:17:06.000000000' '2019-08-07T23:10:23.000000000'
 '2019-08-08T01:22:24.000000000' '2019-08-08T05:52:17.000000000'
 '2019-08-08T13:16:14.000000000' '2019-08-08T20:16:52.000000000'
 '2019-08-10T02:27:26.000000000' '2019-08-10T07:31:31.000000000'
 '2019-08-10T22:08:00.000000000' '2019-08-10T22:51:49.000000000'
 '2019-08-10T23:39:31.000000000' '2019-08-10T23:59:15.000000000'
 '2019-08-11T00:41:26.000000000' '2019-08-11T01:38:37.000000000'
 '2019-08-11T02:28:20.000000000' '2019-08-11T02:59:38.000000000'
 '2019-08-11T04:07:19.000000000' '2019-08-11T04:43:49.000000000'
 '2019-08-11T05:16:01.000000000' '2019-08-11T06:12:21.000000000'
 '2019-08-11T06:50:51.000000000' '2019-08-11T07:31:05.000000000'
 '2019-08-11T10:20:42.000000000' '2019-08-11T15:48:45.000000000'
 '2019-08-11T16:43:42.000000000' '2019-08-11T17:02:09.000000000'
 '2019-08-11T17:39:11.000000000' '2019-08-11T18:58:26.000000000'
 '2019-08-11T22:12:17.000000000' '2019-08-11T22:51:04.000000000'
 '2019-08-11T23:27:28.000000000' '2019-08-12T02:40:46.000000000'
 '2019-08-12T05:38:09.000000000' '2019-08-12T06:39:31.000000000'
 '2019-08-13T13:09:12.000000000' '2019-08-13T14:17:18.000000000'
 '2019-08-13T17:24:18.000000000' '2019-08-13T21:52:25.000000000'
 '2019-08-14T03:53:23.000000000' '2019-08-14T05:15:12.000000000'
 '2019-08-14T06:51:12.000000000' '2019-08-14T07:42:28.000000000'
 '2019-08-14T08:12:45.000000000' '2019-08-14T09:12:31.000000000'
 '2019-08-14T09:33:34.000000000' '2019-08-14T14:16:32.000000000'
 '2019-08-14T15:22:19.000000000' '2019-08-14T18:16:49.000000000'
 '2019-08-15T03:36:05.000000000' '2019-08-15T04:51:27.000000000'
 '2019-08-15T12:02:44.000000000' '2019-08-15T13:16:20.000000000'
 '2019-08-15T14:32:08.000000000' '2019-08-15T15:26:15.000000000'
 '2019-08-15T15:59:52.000000000' '2019-08-15T16:41:41.000000000'
 '2019-08-15T17:31:28.000000000' '2019-08-16T10:24:10.000000000'
 '2019-08-16T11:43:30.000000000' '2019-08-16T13:43:17.000000000'
 '2019-08-16T15:19:26.000000000' '2019-08-16T16:24:37.000000000'
 '2019-08-16T19:28:01.000000000' '2019-08-16T19:57:13.000000000'
 '2019-08-16T20:52:00.000000000' '2019-08-16T21:20:37.000000000'
 '2019-08-16T21:50:44.000000000' '2019-08-16T22:18:17.000000000'
 '2019-08-16T22:52:28.000000000' '2019-08-16T23:24:23.000000000'
 '2019-08-17T00:20:00.000000000' '2019-08-17T01:14:04.000000000'
 '2019-08-17T01:55:02.000000000' '2019-08-17T02:57:31.000000000'
 '2019-08-17T03:59:36.000000000' '2019-08-17T07:03:19.000000000'
 '2019-08-18T08:53:23.000000000' '2019-08-20T01:37:18.000000000'
 '2019-08-20T03:13:17.000000000' '2019-08-20T04:24:27.000000000'
 '2019-08-20T05:17:54.000000000' '2019-08-20T06:23:45.000000000'
 '2019-08-20T07:28:12.000000000' '2019-08-21T00:55:41.000000000'
 '2019-08-21T02:15:29.000000000' '2019-08-21T04:34:37.000000000'
 '2019-08-21T06:17:41.000000000' '2019-08-21T07:13:57.000000000'
 '2019-08-21T08:25:16.000000000' '2019-08-21T17:12:52.000000000'
 '2019-08-22T01:25:48.000000000' '2019-08-22T18:53:21.000000000'
 '2019-08-22T19:54:53.000000000' '2019-08-22T21:36:46.000000000'
 '2019-08-23T09:22:26.000000000' '2019-08-23T09:48:02.000000000'
 '2019-08-23T11:02:27.000000000' '2019-08-24T00:30:35.000000000'
 '2019-08-24T01:29:14.000000000' '2019-08-24T02:36:34.000000000'
 '2019-08-24T03:26:11.000000000' '2019-08-24T04:12:34.000000000'
 '2019-08-24T05:05:23.000000000' '2019-08-24T07:17:34.000000000'
 '2019-08-24T08:04:18.000000000' '2019-08-24T15:33:48.000000000'
 '2019-08-24T17:19:07.000000000' '2019-08-24T18:26:40.000000000'
 '2019-08-24T23:25:49.000000000' '2019-08-25T01:13:12.000000000'
 '2019-08-25T02:48:51.000000000' '2019-08-26T05:43:24.000000000'
 '2019-08-26T15:36:04.000000000' '2019-08-26T22:24:04.000000000'
 '2019-08-30T09:55:11.000000000' '2019-08-30T12:00:51.000000000'
 '2019-08-30T21:42:59.000000000' '2019-08-30T22:05:38.000000000'
 '2019-08-30T23:54:01.000000000' '2019-08-31T00:14:30.000000000'
 '2019-08-31T01:24:06.000000000' '2019-08-31T16:08:47.000000000'
 '2019-08-31T16:28:34.000000000' '2019-08-31T17:10:54.000000000'
 '2019-09-02T06:33:55.000000000' '2019-09-02T07:28:42.000000000'
 '2019-09-02T08:23:04.000000000' '2019-09-02T14:08:54.000000000'
 '2019-09-02T15:03:24.000000000' '2019-09-02T16:09:14.000000000'
 '2019-09-02T16:40:49.000000000' '2019-09-02T17:20:04.000000000'
 '2019-09-02T17:53:45.000000000' '2019-09-04T03:28:13.000000000'
 '2019-09-06T01:11:01.000000000' '2019-09-06T02:22:05.000000000'
 '2019-09-06T04:00:31.000000000' '2019-09-06T05:07:45.000000000'
 '2019-09-06T15:06:00.000000000' '2019-09-06T16:24:28.000000000'
                           'NaT']
Количество уникальных значений: 577
****************************************************************************************************
Уникальные значения столбца 'Bulk 7':
['2019-05-07T18:11:01.000000000' '2019-05-15T22:42:51.000000000'
 '2019-06-17T02:23:02.000000000' '2019-06-23T21:47:17.000000000'
 '2019-07-20T16:22:44.000000000' '2019-07-20T19:54:13.000000000'
 '2019-07-20T23:48:47.000000000' '2019-07-27T03:55:56.000000000'
 '2019-07-27T05:17:30.000000000' '2019-07-27T06:40:27.000000000'
 '2019-07-27T08:58:52.000000000' '2019-07-27T15:19:35.000000000'
 '2019-07-27T18:07:02.000000000' '2019-07-27T19:15:43.000000000'
 '2019-08-12T20:24:56.000000000' '2019-08-12T22:25:23.000000000'
 '2019-08-12T23:46:35.000000000' '2019-08-13T01:36:24.000000000'
 '2019-08-13T04:19:43.000000000' '2019-08-13T07:20:46.000000000'
 '2019-08-13T09:39:09.000000000' '2019-08-13T11:01:00.000000000'
 '2019-09-05T16:39:32.000000000' '2019-09-05T18:40:50.000000000'
 '2019-09-05T19:07:49.000000000'                           'NaT']
Количество уникальных значений: 26
****************************************************************************************************
Уникальные значения столбца 'Bulk 8':
['2019-07-08T17:14:53.000000000'                           'NaT']
Количество уникальных значений: 2
****************************************************************************************************
Уникальные значения столбца 'Bulk 9':
['2019-05-14T11:57:58.000000000' '2019-05-14T12:35:40.000000000'
 '2019-05-14T13:02:53.000000000' '2019-05-14T13:47:34.000000000'
 '2019-05-14T14:10:18.000000000' '2019-05-14T14:51:58.000000000'
 '2019-05-14T15:30:12.000000000' '2019-05-14T16:00:13.000000000'
 '2019-05-14T16:39:17.000000000' '2019-05-14T16:55:09.000000000'
 '2019-07-08T17:23:24.000000000' '2019-07-22T08:33:41.000000000'
 '2019-07-22T10:46:08.000000000' '2019-08-16T05:19:14.000000000'
 '2019-08-16T06:07:48.000000000' '2019-08-16T07:20:13.000000000'
 '2019-08-16T07:52:37.000000000' '2019-08-16T08:40:51.000000000'
 '2019-08-16T09:11:56.000000000'                           'NaT']
Количество уникальных значений: 20
****************************************************************************************************
Уникальные значения столбца 'Bulk 10':
['2019-05-06T07:54:02.000000000' '2019-05-07T01:15:07.000000000'
 '2019-05-07T01:56:53.000000000' '2019-05-07T03:10:27.000000000'
 '2019-05-07T03:32:28.000000000' '2019-05-07T04:36:23.000000000'
 '2019-05-07T05:50:53.000000000' '2019-05-07T06:50:00.000000000'
 '2019-05-14T19:21:25.000000000' '2019-05-14T20:05:17.000000000'
 '2019-05-16T01:56:47.000000000' '2019-05-17T21:07:18.000000000'
 '2019-05-17T21:28:49.000000000' '2019-05-17T22:17:26.000000000'
 '2019-05-17T23:01:27.000000000' '2019-05-17T23:47:29.000000000'
 '2019-05-18T11:18:06.000000000' '2019-05-21T01:58:55.000000000'
 '2019-05-21T03:31:46.000000000' '2019-05-21T04:27:19.000000000'
 '2019-05-21T05:56:02.000000000' '2019-05-21T07:11:02.000000000'
 '2019-05-21T08:19:09.000000000' '2019-05-21T09:07:35.000000000'
 '2019-05-21T09:58:45.000000000' '2019-05-21T21:36:58.000000000'
 '2019-05-21T22:37:19.000000000' '2019-05-22T08:27:23.000000000'
 '2019-05-22T09:24:11.000000000' '2019-05-22T10:18:51.000000000'
 '2019-05-22T12:14:58.000000000' '2019-05-22T12:46:08.000000000'
 '2019-05-24T18:28:15.000000000' '2019-05-24T19:12:02.000000000'
 '2019-05-24T19:38:09.000000000' '2019-05-24T20:30:00.000000000'
 '2019-05-26T12:19:31.000000000' '2019-05-26T13:04:03.000000000'
 '2019-05-28T20:36:47.000000000' '2019-05-30T00:35:25.000000000'
 '2019-05-31T19:29:36.000000000' '2019-06-01T00:59:53.000000000'
 '2019-06-01T10:09:47.000000000' '2019-06-02T03:06:59.000000000'
 '2019-06-02T05:53:41.000000000' '2019-06-02T06:26:35.000000000'
 '2019-06-02T07:27:58.000000000' '2019-06-02T08:24:42.000000000'
 '2019-06-02T11:20:31.000000000' '2019-06-02T14:51:45.000000000'
 '2019-06-06T16:01:44.000000000' '2019-06-10T20:08:17.000000000'
 '2019-06-10T20:43:32.000000000' '2019-06-10T22:58:03.000000000'
 '2019-06-13T02:15:11.000000000' '2019-06-13T04:15:09.000000000'
 '2019-06-13T05:22:48.000000000' '2019-06-20T05:41:49.000000000'
 '2019-06-20T06:37:50.000000000' '2019-06-20T18:05:33.000000000'
 '2019-06-20T20:54:25.000000000' '2019-06-20T21:59:35.000000000'
 '2019-06-23T12:48:51.000000000' '2019-06-23T14:22:38.000000000'
 '2019-06-25T19:42:29.000000000' '2019-06-25T20:06:18.000000000'
 '2019-06-25T21:41:50.000000000' '2019-06-25T22:14:55.000000000'
 '2019-06-27T21:50:53.000000000' '2019-06-27T22:21:02.000000000'
 '2019-06-27T23:04:38.000000000' '2019-06-27T23:28:03.000000000'
 '2019-06-28T00:30:14.000000000' '2019-06-28T15:13:56.000000000'
 '2019-06-28T16:19:20.000000000' '2019-06-28T17:19:09.000000000'
 '2019-06-28T17:47:14.000000000' '2019-06-28T18:23:07.000000000'
 '2019-06-28T19:01:44.000000000' '2019-06-30T21:21:30.000000000'
 '2019-07-01T01:12:27.000000000' '2019-07-01T02:45:36.000000000'
 '2019-07-01T05:20:39.000000000' '2019-07-04T12:28:56.000000000'
 '2019-07-06T04:50:19.000000000' '2019-07-06T05:31:18.000000000'
 '2019-07-06T06:27:50.000000000' '2019-07-06T06:41:52.000000000'
 '2019-07-06T07:28:48.000000000' '2019-07-06T09:23:17.000000000'
 '2019-07-06T09:57:49.000000000' '2019-07-06T10:39:52.000000000'
 '2019-07-09T05:29:31.000000000' '2019-07-09T06:19:17.000000000'
 '2019-07-09T06:52:55.000000000' '2019-07-09T20:46:56.000000000'
 '2019-07-09T21:37:00.000000000' '2019-07-12T21:47:41.000000000'
 '2019-07-18T19:49:54.000000000' '2019-07-18T21:09:10.000000000'
 '2019-07-21T21:40:41.000000000' '2019-07-21T22:29:14.000000000'
 '2019-07-24T06:19:03.000000000' '2019-07-28T07:49:06.000000000'
 '2019-07-28T08:32:44.000000000' '2019-07-28T08:49:53.000000000'
 '2019-07-28T09:44:48.000000000' '2019-07-28T10:36:29.000000000'
 '2019-07-28T11:06:33.000000000' '2019-07-28T11:56:38.000000000'
 '2019-07-28T13:24:37.000000000' '2019-07-28T14:06:25.000000000'
 '2019-07-29T08:36:32.000000000' '2019-07-29T10:24:32.000000000'
 '2019-07-29T11:14:34.000000000' '2019-07-29T11:34:33.000000000'
 '2019-07-29T12:15:00.000000000' '2019-07-29T12:47:14.000000000'
 '2019-07-30T10:48:38.000000000' '2019-07-30T13:36:57.000000000'
 '2019-07-31T11:10:59.000000000' '2019-07-31T12:22:40.000000000'
 '2019-07-31T12:44:11.000000000' '2019-07-31T14:56:10.000000000'
 '2019-08-01T20:17:21.000000000' '2019-08-01T20:35:37.000000000'
 '2019-08-01T23:49:43.000000000' '2019-08-02T00:25:09.000000000'
 '2019-08-02T02:48:08.000000000' '2019-08-02T03:17:23.000000000'
 '2019-08-03T22:18:02.000000000' '2019-08-04T02:06:48.000000000'
 '2019-08-04T03:13:59.000000000' '2019-08-04T05:57:13.000000000'
 '2019-08-04T16:34:19.000000000' '2019-08-04T19:55:58.000000000'
 '2019-08-04T21:00:53.000000000' '2019-08-04T22:48:58.000000000'
 '2019-08-04T23:12:11.000000000' '2019-08-07T04:55:13.000000000'
 '2019-08-10T11:31:40.000000000' '2019-08-10T12:17:52.000000000'
 '2019-08-10T13:21:18.000000000' '2019-08-14T23:20:47.000000000'
 '2019-08-15T00:05:05.000000000' '2019-08-15T01:09:17.000000000'
 '2019-08-15T01:39:57.000000000' '2019-08-15T20:10:01.000000000'
 '2019-08-20T17:15:15.000000000' '2019-08-20T19:51:47.000000000'
 '2019-08-20T20:29:52.000000000' '2019-08-20T21:07:11.000000000'
 '2019-08-20T22:02:14.000000000' '2019-08-20T22:25:23.000000000'
 '2019-08-26T07:24:11.000000000' '2019-08-27T02:07:02.000000000'
 '2019-08-28T15:35:38.000000000' '2019-08-28T15:53:05.000000000'
 '2019-08-28T16:48:18.000000000' '2019-08-28T18:03:47.000000000'
 '2019-08-29T07:01:20.000000000' '2019-08-29T08:35:38.000000000'
 '2019-08-29T13:18:33.000000000' '2019-08-29T14:04:23.000000000'
 '2019-08-29T14:17:57.000000000' '2019-08-31T04:56:48.000000000'
 '2019-08-31T05:55:31.000000000' '2019-08-31T06:37:35.000000000'
 '2019-08-31T07:28:39.000000000' '2019-09-02T03:21:10.000000000'
 '2019-09-02T03:32:49.000000000' '2019-09-02T04:45:41.000000000'
 '2019-09-03T05:40:49.000000000' '2019-09-03T05:57:25.000000000'
 '2019-09-04T09:44:31.000000000' '2019-09-04T10:03:22.000000000'
                           'NaT']
Количество уникальных значений: 177
****************************************************************************************************
Уникальные значения столбца 'Bulk 11':
['2019-05-05T23:43:24.000000000' '2019-05-06T16:18:00.000000000'
 '2019-05-07T00:01:30.000000000' '2019-05-07T09:53:19.000000000'
 '2019-05-07T18:04:04.000000000' '2019-05-08T15:39:24.000000000'
 '2019-05-09T08:58:16.000000000' '2019-05-09T14:23:12.000000000'
 '2019-05-10T07:53:26.000000000' '2019-05-10T14:46:18.000000000'
 '2019-05-10T18:36:30.000000000' '2019-05-10T21:21:48.000000000'
 '2019-05-17T06:42:42.000000000' '2019-05-17T11:14:08.000000000'
 '2019-05-18T23:16:30.000000000' '2019-05-18T23:55:19.000000000'
 '2019-05-21T17:57:26.000000000' '2019-05-23T03:31:40.000000000'
 '2019-05-23T11:26:07.000000000' '2019-05-24T09:36:26.000000000'
 '2019-05-25T06:37:43.000000000' '2019-05-25T13:54:57.000000000'
 '2019-05-27T13:16:13.000000000' '2019-06-04T02:54:36.000000000'
 '2019-06-04T04:21:50.000000000' '2019-06-05T12:18:07.000000000'
 '2019-06-05T23:27:56.000000000' '2019-06-06T11:43:48.000000000'
 '2019-06-07T08:42:42.000000000' '2019-06-07T15:41:43.000000000'
 '2019-06-08T00:04:16.000000000' '2019-06-11T07:22:27.000000000'
 '2019-06-11T19:15:50.000000000' '2019-06-11T23:01:45.000000000'
 '2019-06-12T01:27:39.000000000' '2019-06-12T02:31:44.000000000'
 '2019-06-12T05:40:52.000000000' '2019-06-12T17:19:45.000000000'
 '2019-06-12T18:03:07.000000000' '2019-06-12T19:34:53.000000000'
 '2019-06-12T21:30:38.000000000' '2019-06-13T14:44:39.000000000'
 '2019-06-14T02:25:19.000000000' '2019-06-14T13:30:53.000000000'
 '2019-06-14T20:24:48.000000000' '2019-06-15T12:48:22.000000000'
 '2019-06-15T17:46:40.000000000' '2019-06-17T09:00:16.000000000'
 '2019-06-17T10:42:37.000000000' '2019-06-22T00:44:00.000000000'
 '2019-06-25T15:04:33.000000000' '2019-06-25T18:13:24.000000000'
 '2019-06-27T10:50:36.000000000' '2019-06-29T08:23:23.000000000'
 '2019-06-29T14:55:02.000000000' '2019-07-02T01:14:00.000000000'
 '2019-07-04T11:37:11.000000000' '2019-07-05T16:23:29.000000000'
 '2019-07-05T16:46:22.000000000' '2019-07-07T07:56:08.000000000'
 '2019-07-09T04:40:52.000000000' '2019-07-10T04:55:12.000000000'
 '2019-07-11T05:54:32.000000000' '2019-07-11T09:25:45.000000000'
 '2019-07-12T03:15:38.000000000' '2019-07-12T09:06:13.000000000'
 '2019-07-12T13:43:25.000000000' '2019-07-13T05:33:18.000000000'
 '2019-07-19T07:43:27.000000000' '2019-07-19T16:46:45.000000000'
 '2019-07-19T18:55:00.000000000' '2019-07-19T20:59:43.000000000'
 '2019-07-19T23:15:16.000000000' '2019-07-20T01:48:13.000000000'
 '2019-07-20T04:22:08.000000000' '2019-07-20T09:56:49.000000000'
 '2019-07-20T15:29:10.000000000' '2019-07-21T00:22:33.000000000'
 '2019-07-22T00:46:22.000000000' '2019-07-22T01:40:34.000000000'
 '2019-07-22T03:16:46.000000000' '2019-07-22T04:01:49.000000000'
 '2019-07-22T17:27:34.000000000' '2019-07-23T05:11:26.000000000'
 '2019-07-24T07:33:14.000000000' '2019-07-24T07:57:08.000000000'
 '2019-07-24T09:27:39.000000000' '2019-07-24T10:34:50.000000000'
 '2019-07-24T11:17:22.000000000' '2019-07-24T11:57:35.000000000'
 '2019-07-24T13:00:50.000000000' '2019-07-24T13:50:39.000000000'
 '2019-07-24T15:43:34.000000000' '2019-07-24T21:07:51.000000000'
 '2019-07-24T23:08:34.000000000' '2019-07-24T23:47:11.000000000'
 '2019-07-26T00:20:40.000000000' '2019-07-26T20:00:10.000000000'
 '2019-07-27T00:04:49.000000000' '2019-07-28T03:17:47.000000000'
 '2019-07-28T06:11:49.000000000' '2019-07-29T03:10:59.000000000'
 '2019-07-29T22:45:53.000000000' '2019-07-30T14:02:12.000000000'
 '2019-07-30T23:01:15.000000000' '2019-07-31T03:48:25.000000000'
 '2019-07-31T04:30:00.000000000' '2019-07-31T05:07:02.000000000'
 '2019-07-31T05:40:37.000000000' '2019-07-31T13:38:49.000000000'
 '2019-08-01T01:51:27.000000000' '2019-08-02T04:32:32.000000000'
 '2019-08-03T03:12:11.000000000' '2019-08-03T11:37:16.000000000'
 '2019-08-03T11:48:52.000000000' '2019-08-03T13:07:21.000000000'
 '2019-08-04T18:37:48.000000000' '2019-08-05T10:10:45.000000000'
 '2019-08-05T11:31:20.000000000' '2019-08-07T17:36:24.000000000'
 '2019-08-07T18:45:21.000000000' '2019-08-07T22:49:01.000000000'
 '2019-08-07T23:41:22.000000000' '2019-08-08T01:22:24.000000000'
 '2019-08-08T13:09:34.000000000' '2019-08-08T21:40:01.000000000'
 '2019-08-09T02:24:16.000000000' '2019-08-09T16:12:37.000000000'
 '2019-08-10T02:22:53.000000000' '2019-08-10T07:31:31.000000000'
 '2019-08-11T03:14:52.000000000' '2019-08-11T04:38:50.000000000'
 '2019-08-11T05:21:01.000000000' '2019-08-11T06:04:28.000000000'
 '2019-08-11T16:27:29.000000000' '2019-08-11T22:46:02.000000000'
 '2019-08-12T03:20:19.000000000' '2019-08-12T05:35:47.000000000'
 '2019-08-12T06:52:27.000000000' '2019-08-12T10:18:47.000000000'
 '2019-08-13T14:10:57.000000000' '2019-08-13T17:18:52.000000000'
 '2019-08-13T19:39:29.000000000' '2019-08-13T20:51:12.000000000'
 '2019-08-13T21:26:01.000000000' '2019-08-13T22:09:13.000000000'
 '2019-08-14T08:27:26.000000000' '2019-08-14T14:10:22.000000000'
 '2019-08-14T17:53:38.000000000' '2019-08-15T03:24:52.000000000'
 '2019-08-15T04:39:27.000000000' '2019-08-16T11:49:05.000000000'
 '2019-08-16T13:36:53.000000000' '2019-08-16T22:25:25.000000000'
 '2019-08-16T23:24:23.000000000' '2019-08-17T01:16:25.000000000'
 '2019-08-17T01:45:04.000000000' '2019-08-17T03:49:17.000000000'
 '2019-08-17T07:03:19.000000000' '2019-08-18T05:07:43.000000000'
 '2019-08-21T04:28:31.000000000' '2019-08-21T06:02:28.000000000'
 '2019-08-21T07:13:57.000000000' '2019-08-22T00:59:25.000000000'
 '2019-08-22T19:36:51.000000000' '2019-08-22T20:11:37.000000000'
 '2019-08-24T06:42:47.000000000' '2019-08-24T07:13:12.000000000'
 '2019-08-24T08:04:18.000000000' '2019-08-24T17:19:07.000000000'
 '2019-08-30T23:12:49.000000000' '2019-08-31T00:22:35.000000000'
 '2019-08-31T01:24:06.000000000' '2019-09-06T01:04:04.000000000'
 '2019-09-06T02:22:05.000000000' '2019-09-06T03:54:36.000000000'
 '2019-09-06T05:03:14.000000000'                           'NaT']
Количество уникальных значений: 178
****************************************************************************************************
Уникальные значения столбца 'Bulk 12':
['2019-05-03T11:24:31.000000000' '2019-05-03T11:53:30.000000000'
 '2019-05-03T12:27:13.000000000' ... '2019-09-06T12:18:35.000000000'
 '2019-09-06T15:01:44.000000000'                           'NaT']
Количество уникальных значений: 2451
****************************************************************************************************
Уникальные значения столбца 'Bulk 13':
['2019-05-05T02:10:21.000000000' '2019-05-11T03:35:36.000000000'
 '2019-05-11T08:18:48.000000000' '2019-05-16T05:29:33.000000000'
 '2019-05-28T08:15:29.000000000' '2019-06-01T18:07:38.000000000'
 '2019-06-09T08:18:22.000000000' '2019-06-21T20:10:19.000000000'
 '2019-06-30T14:38:12.000000000' '2019-07-02T11:21:01.000000000'
 '2019-07-05T14:20:41.000000000' '2019-07-11T03:10:43.000000000'
 '2019-07-29T02:58:13.000000000' '2019-08-07T06:23:16.000000000'
 '2019-08-22T14:48:22.000000000' '2019-08-26T07:00:42.000000000'
 '2019-08-29T14:49:47.000000000' '2019-09-01T01:53:02.000000000'
                           'NaT']
Количество уникальных значений: 19
****************************************************************************************************
Уникальные значения столбца 'Bulk 14':
['2019-05-03T11:14:50.000000000' '2019-05-03T11:48:37.000000000'
 '2019-05-03T12:21:01.000000000' ... '2019-09-06T16:07:29.000000000'
 '2019-09-06T17:26:33.000000000'                           'NaT']
Количество уникальных значений: 2807
****************************************************************************************************
Уникальные значения столбца 'Bulk 15':
['2019-05-03T11:10:43.000000000' '2019-05-03T11:44:39.000000000'
 '2019-05-03T12:16:16.000000000' ... '2019-09-06T16:01:34.000000000'
 '2019-09-06T17:23:15.000000000'                           'NaT']
Количество уникальных значений: 2249
****************************************************************************************************


df_bulk_time


df_bulk_time['start_time_bulk'] = df_bulk_time.min(axis=1)
df_bulk_time['finish_time_bulk'] = df_bulk_time.max(axis=1)
df_bulk_time['duration_bulk'] = (
    df_bulk_time['finish_time_bulk'] - df_bulk_time['start_time_bulk']).astype('timedelta64[s]')


df_bulk_time


df_bulk_time = df_bulk_time[['start_time_bulk', 'finish_time_bulk','duration_bulk']]
df_bulk_time


# Сводная статистика для всех столбцов в `df_bulk_time`
df_bulk_time.describe(include='all', datetime_is_numeric=True)


# Создаем графическую фигуру
plt.figure(figsize=(15, 8))
# Pазбиваем графическую фигуру на 2 графика
# Первый график
plt.subplot(2, 1, 1)
# Гистограмма для 'duration_bulk'
sns.histplot(data=df_bulk_time['duration_bulk'],
             alpha=0.8,
             label='Длительность обработки с применением сыпучих материалов(с)')

plt.minorticks_on()
plt.grid(which='minor',
         linestyle=':')
plt.grid(True)
plt.xlabel('')
plt.ylabel('Частота',
           fontsize=12,
           color='DarkSlateGray')
plt.legend()

# Второй график
plt.subplot(2, 1, 2)
# Боксплот для 'duration_bulk' 
sns.boxplot(data=df_bulk_time,
            x='duration_bulk',
            orient='horizontal')

plt.minorticks_on()
plt.grid(which='minor',
         linestyle=':')
plt.grid(True)
plt.xlabel('Длительность обработки с применением сыпучих материалов',
           fontsize=12,
           color='DarkSlateGray')
plt.ylabel('')
plt.suptitle('Распределение длительность обработки с применением сыпучих материалов (с)',
             fontsize=15,
             color='DarkSlateGray')
plt.show()


# Посмотрим уникальные значения по колонкам `df_bulk_time`
get_unique_sorted(df_bulk_time)

Уникальные значения столбца 'start_time_bulk':
['2019-05-03T11:10:43.000000000' '2019-05-03T11:36:50.000000000'
 '2019-05-03T12:16:16.000000000' ... '2019-09-06T14:48:06.000000000'
 '2019-09-06T16:01:34.000000000' '2019-09-06T17:23:15.000000000']
Количество уникальных значений: 3129
****************************************************************************************************
Уникальные значения столбца 'finish_time_bulk':
['2019-05-03T11:28:48.000000000' '2019-05-03T11:53:30.000000000'
 '2019-05-03T12:32:39.000000000' ... '2019-09-06T15:06:00.000000000'
 '2019-09-06T16:24:28.000000000' '2019-09-06T17:26:33.000000000']
Количество уникальных значений: 3129
****************************************************************************************************
Уникальные значения столбца 'duration_bulk':
[    0.    82.   123. ...  9336. 11552. 13683.]
Количество уникальных значений: 1528
****************************************************************************************************


# Создаем графическую фигуру
plt.figure(figsize=(15, 8))
# Pазбиваем графическую фигуру на 2 графика
# Первый график
plt.subplot(2, 1, 1)
sns.histplot(data=df_bulk_time[df_bulk_time['duration_bulk'] != 0]['duration_bulk'],
             alpha=0.8,
             label='Длительность обработки с применением сыпучих материалов (с)')

plt.minorticks_on()
plt.grid(which='minor',
         linestyle=':')
plt.grid(True)
plt.xlabel('')
plt.ylabel('Частота',
           fontsize=12,
           color='DarkSlateGray')
plt.legend()

# Второй график
plt.subplot(2, 1, 2)
# Боксплот для 'duration_bulk' 
sns.boxplot(data=df_bulk_time[df_bulk_time['duration_bulk'] != 0],
            x='duration_bulk',
            orient='horizontal')

plt.minorticks_on()
plt.grid(which='minor',
         linestyle=':')
plt.grid(True)
plt.xlabel('Длительность обработки с применением сыпучих материалов (с)',
           fontsize=12,
           color='DarkSlateGray')
plt.ylabel('')
plt.suptitle('Распределение длительности обработки с применением сыпучих материалов (с)',
             fontsize=15,
             color='DarkSlateGray')
plt.show()


# Проверим столбец `key` на уникальность значений
df_gas['key'].duplicated().sum()

0


# Столбец `key` установим в качестве индекса датафрейма
df_gas = df_gas.set_index('key')


# Посмотрим уникальные значения по колонкам `df_gas`
get_unique_sorted(df_gas)

Уникальные значения столбца 'Газ 1':
[8.39852910e-03 1.66956024e-02 2.63028954e-01 ... 5.21423726e+01
 6.09356892e+01 7.79950397e+01]
Количество уникальных значений: 3239
****************************************************************************************************


df_gas


# Распределение количества газа
# Создаем графическую фигуру
plt.figure(figsize=(15, 8))
# Pазбиваем графическую фигуру на 2 графика
# Первый график
plt.subplot(2, 1, 1)
# Гистограмма для 'Газ 1'
sns.histplot(data=df_gas['Газ 1'],
             alpha=0.8,
             label='Количество газа')

plt.minorticks_on()
plt.grid(which='minor',
         linestyle=':')
plt.grid(True)
plt.xlabel('')
plt.ylabel('Частота',
           fontsize=12,
           color='DarkSlateGray')
plt.legend()

# Второй график
plt.subplot(2, 1, 2)
# Боксплот для 'Газ 1'
sns.boxplot(data=df_gas,
            x='Газ 1',
            orient='horizontal')

plt.minorticks_on()
plt.grid(which='minor',
         linestyle=':')
plt.grid(True)
plt.xlabel('Количество газа',
           fontsize=12,
           color='DarkSlateGray')
plt.ylabel('')
plt.suptitle('Распределение количества газа',
             fontsize=15,
             color='DarkSlateGray')
plt.show()

# Сводная статистика для всех столбцов в `df_gas`
df_gas.describe()


# Приведём таблицу к общепринятому виду
df_gas = df_gas.rename(columns={'Газ 1':'gas_quantities'})
df_gas


# Проверим столбец `key` на уникальность значений
df_temp['key'].duplicated().sum()

14876


# Посмотрим уникальные значения по колонкам `df_temp`
get_unique_sorted(df_temp)

Уникальные значения столбца 'key':
[   1    2    3 ... 3239 3240 3241]
Количество уникальных значений: 3216
****************************************************************************************************
Уникальные значения столбца 'Время замера':
['2019-05-03T11:02:04.000000000' '2019-05-03T11:07:18.000000000'
 '2019-05-03T11:11:34.000000000' ... '2019-09-06T17:21:48.000000000'
 '2019-09-06T17:24:44.000000000' '2019-09-06T17:30:05.000000000']
Количество уникальных значений: 18092
****************************************************************************************************
Уникальные значения столбца 'Температура':
[1191. 1204. 1208. 1218. 1227. 1515. 1519. 1520. 1521. 1522. 1525. 1526.
 1527. 1528. 1529. 1530. 1531. 1532. 1533. 1534. 1535. 1536. 1537. 1538.
 1539. 1540. 1541. 1542. 1543. 1544. 1545. 1546. 1547. 1548. 1549. 1550.
 1551. 1552. 1553. 1554. 1555. 1556. 1557. 1558. 1559. 1560. 1561. 1562.
 1563. 1564. 1565. 1566. 1567. 1568. 1569. 1570. 1571. 1572. 1573. 1574.
 1575. 1576. 1577. 1578. 1579. 1580. 1581. 1582. 1583. 1584. 1585. 1586.
 1587. 1588. 1589. 1590. 1591. 1592. 1593. 1594. 1595. 1596. 1597. 1598.
 1599. 1600. 1601. 1602. 1603. 1604. 1605. 1606. 1607. 1608. 1609. 1610.
 1611. 1612. 1613. 1614. 1615. 1616. 1617. 1618. 1619. 1620. 1621. 1622.
 1623. 1624. 1625. 1626. 1627. 1628. 1629. 1630. 1631. 1632. 1633. 1634.
 1635. 1636. 1637. 1638. 1639. 1640. 1641. 1642. 1643. 1644. 1645. 1646.
 1647. 1648. 1649. 1650. 1651. 1652. 1653. 1654. 1655. 1656. 1657. 1658.
 1659. 1660. 1661. 1662. 1663. 1665. 1666. 1667. 1668. 1669. 1670. 1671.
 1672. 1673. 1674. 1675. 1676. 1678. 1679. 1680. 1681. 1684. 1690. 1691.
 1696. 1700. 1704. 1705.   nan]
Количество уникальных значений: 173
****************************************************************************************************


# Сводная статистика для всех столбцов в `df_temp`
df_temp.describe(include='all', datetime_is_numeric=True)


# Распределение количества газа
# Создаем графическую фигуру
plt.figure(figsize=(15, 8))
# Pазбиваем графическую фигуру на 2 графика
# Первый график
plt.subplot(2, 1, 1)
# Гистограмма для 'Температура'
sns.histplot(data=df_temp['Температура'],
             alpha=0.8,
             label='Температура')

plt.minorticks_on()
plt.grid(which='minor',
         linestyle=':')
plt.grid(True)
plt.xlabel('')
plt.ylabel('Частота',
           fontsize=12,
           color='DarkSlateGray')
plt.legend()

# Второй график
plt.subplot(2, 1, 2)
# Боксплот для 'Температура'
sns.boxplot(data=df_temp,
            x='Температура',
            orient='horizontal')

plt.minorticks_on()
plt.grid(which='minor',
         linestyle=':')
plt.grid(True)
plt.xlabel('Температура',
           fontsize=12,
           color='DarkSlateGray')
plt.ylabel('')
plt.suptitle('Распределение показателей температуры',
             fontsize=15,
             color='DarkSlateGray')
plt.show()


df_temp


# Распределение партий по количеству замеров температур
df_count = df_temp.groupby('key')['key'].count().value_counts().reset_index()
df_count.columns = ['кол-во замеров', 'кол-во партий']
total_batches = df_count['кол-во партий'].sum()
df_count['% кол-ва партий'] = df_count['кол-во партий'] / total_batches * 100
df_count


# Удаляем строки, где количество замеров меньше двух и температура NaN
df_temp = df_temp.dropna(subset=['Температура'])
df_temp = df_temp.groupby('key').filter(lambda x: len(x) >= 2)
df_temp


# Группируем данные по key
grouped_data = df_temp.groupby('key')

# Собираем данные по требуемым колонкам
df_result_temp = pd.DataFrame()
df_result_temp['time_first_measurement'] = grouped_data['Время замера'].first()
df_result_temp['time_last_measurement'] = grouped_data['Время замера'].last()
df_result_temp['temperature_first_measurement'] = grouped_data['Температура'].first()
df_result_temp['temperature_last_measurement'] = grouped_data['Температура'].last()
df_result_temp['time_between_measurements'] = (df_result_temp['time_last_measurement'] - \
    df_result_temp['time_first_measurement']).astype('timedelta64[s]')

# Выводим результат
df_result_temp


# Создаем графическую фигуру
plt.figure(figsize=(15, 8))
# Pазбиваем графическую фигуру на 2 графика
# Первый график
plt.subplot(2, 1, 1)
# Гистограмма для 'temperature_first_measurement'
sns.histplot(data=df_result_temp['temperature_first_measurement'],
             color='green',
             alpha=0.8,
             label='Температура при первом замере')
# Гистограмма для 'temperature_last_measurement'
sns.histplot(data=df_result_temp['temperature_last_measurement'],
             color='red',
             alpha=0.5,
             label='Температура при последнем замере')

plt.minorticks_on()
plt.grid(which='minor',
         linestyle=':')
plt.grid(True)
plt.xlabel('')
plt.ylabel('Частота',
           fontsize=12,
           color='DarkSlateGray')
plt.legend()

# Второй график
plt.subplot(2, 1, 2)
# Боксплот для 'temperature_first_measurement' и 'temperature_last_measurement'
sns.boxplot(data=df_result_temp[['temperature_first_measurement', 'temperature_last_measurement']],
            orient='horizontal',
            palette=['green', 'red'])

plt.minorticks_on()
plt.grid(which='minor',
         linestyle=':')
plt.grid(True)
plt.xlabel('Температура',
           fontsize=12,
           color='DarkSlateGray')
plt.ylabel('')
plt.suptitle('Распределение температуры (при первом и при последнем замере)',
             fontsize=15,
             color='DarkSlateGray')
plt.show()


df_result_temp[(df_result_temp['temperature_first_measurement'] < 1300) | (
    df_result_temp['temperature_last_measurement'] < 1300)]


# Проверим столбец `key` на уникальность значений
df_wire['key'].duplicated().sum()

0


# Столбец `key` установим в качестве индекса датафрейма
df_wire = df_wire.set_index('key')


# Сводная статистика для всех столбцов в `df_wire`
df_wire.describe()


# Посмотрим на уникальную запись с добавлением `Wire 5`
df_wire[df_wire['Wire 5'].notna()]


# Визуализируем распределения по колонкам `df_wire`
plt.figure(figsize=(15, 10))
sns.boxplot(data=df_wire,
            orient='horizontal')

plt.xlabel('Объем подачи проволочных материалов',
           fontsize=12,
           color='DarkSlateGray')
plt.ylabel('Материал',
           fontsize=12,
           color='DarkSlateGray')
plt.title('Boxplot подачи проволочных материалов',
          fontsize=15,
          color='DarkSlateGray')

plt.minorticks_on()
plt.grid(which='minor',
         linestyle=':')
plt.grid(True)


# Заполним значения NaN в DataFrame нулями
df_wire = df_wire.fillna(0)
# Рассчитаем сумму каждой строки и создим новый столбец с названием `wire_sum`
df_wire['wire_sum'] = df_wire.sum(axis=1)
# Заменим пробелы в названиях столбцов на нижнее подчеркивание и приведем их к нижнему регистру
df_wire.columns = [x.replace(' ','_').lower() for x in  df_wire.columns]
df_wire


# Проверим столбец `key` на уникальность значений
df_wire_time['key'].duplicated().sum()

0


# Столбец `key` установим в качестве индекса датафрейма
df_wire_time = df_wire_time.set_index('key')


# Посмотрим уникальные значения по колонкам `df_wire_time`
get_unique_sorted(df_wire_time)

Уникальные значения столбца 'Wire 1':
['2019-05-03T11:06:19.000000000' '2019-05-03T11:36:50.000000000'
 '2019-05-03T12:11:46.000000000' ... '2019-09-06T15:33:55.000000000'
 '2019-09-06T17:10:06.000000000'                           'NaT']
Количество уникальных значений: 3056
****************************************************************************************************
Уникальные значения столбца 'Wire 2':
['2019-05-03T13:15:34.000000000' '2019-05-03T13:48:52.000000000'
 '2019-05-03T15:39:37.000000000' ... '2019-09-06T01:38:22.000000000'
 '2019-09-06T07:35:40.000000000'                           'NaT']
Количество уникальных значений: 1080
****************************************************************************************************
Уникальные значения столбца 'Wire 3':
['2019-05-04T04:34:27.000000000' '2019-05-04T05:41:29.000000000'
 '2019-05-07T15:39:35.000000000' '2019-05-07T17:12:49.000000000'
 '2019-05-07T22:24:56.000000000' '2019-05-07T23:37:44.000000000'
 '2019-05-08T02:25:41.000000000' '2019-05-08T06:25:36.000000000'
 '2019-05-08T07:00:18.000000000' '2019-05-08T07:47:18.000000000'
 '2019-05-19T02:20:39.000000000' '2019-05-19T02:57:29.000000000'
 '2019-05-19T10:05:03.000000000' '2019-05-19T11:07:37.000000000'
 '2019-06-04T12:35:52.000000000' '2019-06-10T15:07:04.000000000'
 '2019-06-12T13:28:13.000000000' '2019-06-12T14:01:36.000000000'
 '2019-06-12T14:38:14.000000000' '2019-06-14T02:18:05.000000000'
 '2019-06-14T21:33:28.000000000' '2019-06-14T22:16:16.000000000'
 '2019-07-10T18:21:05.000000000' '2019-07-10T19:02:13.000000000'
 '2019-07-10T20:08:55.000000000' '2019-07-11T18:01:03.000000000'
 '2019-07-11T19:03:45.000000000' '2019-07-19T04:01:03.000000000'
 '2019-07-19T05:01:12.000000000' '2019-07-19T05:51:49.000000000'
 '2019-07-21T04:23:54.000000000' '2019-07-21T10:04:47.000000000'
 '2019-07-27T03:44:01.000000000' '2019-07-27T05:11:06.000000000'
 '2019-07-27T06:26:25.000000000' '2019-07-27T08:51:53.000000000'
 '2019-07-27T14:05:34.000000000' '2019-07-27T17:11:04.000000000'
 '2019-07-27T20:02:21.000000000' '2019-08-10T14:44:43.000000000'
 '2019-08-10T15:43:16.000000000' '2019-08-10T19:15:17.000000000'
 '2019-08-10T20:02:36.000000000' '2019-08-10T20:53:12.000000000'
 '2019-08-11T03:33:25.000000000' '2019-08-12T20:12:36.000000000'
 '2019-08-12T22:14:01.000000000' '2019-08-12T23:35:31.000000000'
 '2019-08-13T01:29:08.000000000' '2019-08-13T03:21:46.000000000'
 '2019-08-13T07:04:38.000000000' '2019-08-13T07:52:42.000000000'
 '2019-08-13T09:36:56.000000000' '2019-08-17T09:26:28.000000000'
 '2019-08-17T10:16:32.000000000' '2019-08-17T10:55:55.000000000'
 '2019-08-25T23:47:50.000000000' '2019-08-26T01:15:36.000000000'
 '2019-08-26T02:15:46.000000000' '2019-08-26T03:34:51.000000000'
 '2019-09-01T12:41:21.000000000' '2019-09-01T13:28:06.000000000'
 '2019-09-02T07:14:44.000000000'                           'NaT']
Количество уникальных значений: 64
****************************************************************************************************
Уникальные значения столбца 'Wire 4':
['2019-05-07T15:19:17.000000000' '2019-05-07T16:46:56.000000000'
 '2019-05-07T23:21:42.000000000' '2019-07-20T16:11:26.000000000'
 '2019-07-20T23:21:17.000000000' '2019-07-27T03:25:58.000000000'
 '2019-07-27T05:07:46.000000000' '2019-07-27T06:23:07.000000000'
 '2019-07-27T08:32:14.000000000' '2019-07-27T13:57:58.000000000'
 '2019-07-27T17:07:27.000000000' '2019-07-27T19:41:36.000000000'
 '2019-08-12T19:51:09.000000000' '2019-08-13T03:16:45.000000000'
                           'NaT']
Количество уникальных значений: 15
****************************************************************************************************
Уникальные значения столбца 'Wire 5':
['2019-08-13T06:14:30.000000000'                           'NaT']
Количество уникальных значений: 2
****************************************************************************************************
Уникальные значения столбца 'Wire 6':
['2019-05-07T14:46:05.000000000' '2019-05-07T16:16:34.000000000'
 '2019-05-07T17:37:05.000000000' '2019-05-07T21:53:14.000000000'
 '2019-05-07T23:01:42.000000000' '2019-05-08T01:53:35.000000000'
 '2019-05-08T06:19:00.000000000' '2019-05-08T06:54:07.000000000'
 '2019-05-08T07:39:19.000000000' '2019-05-08T09:01:22.000000000'
 '2019-05-08T14:00:10.000000000' '2019-05-08T15:07:48.000000000'
 '2019-05-08T17:02:28.000000000' '2019-05-08T17:52:46.000000000'
 '2019-05-08T18:38:10.000000000' '2019-05-08T19:14:44.000000000'
 '2019-05-08T20:11:04.000000000' '2019-05-08T21:09:25.000000000'
 '2019-05-08T21:47:30.000000000' '2019-05-08T22:48:56.000000000'
 '2019-05-08T23:48:40.000000000' '2019-05-09T00:31:45.000000000'
 '2019-05-09T02:52:53.000000000' '2019-05-09T03:42:13.000000000'
 '2019-05-09T05:40:27.000000000' '2019-07-27T02:13:58.000000000'
 '2019-07-27T03:11:17.000000000' '2019-07-27T04:47:06.000000000'
 '2019-07-27T06:09:34.000000000' '2019-07-27T08:24:44.000000000'
 '2019-07-27T13:48:37.000000000' '2019-07-27T16:40:06.000000000'
 '2019-07-27T19:15:43.000000000' '2019-07-27T22:52:21.000000000'
 '2019-07-28T01:07:46.000000000' '2019-07-28T01:40:36.000000000'
 '2019-07-28T05:00:32.000000000' '2019-07-28T10:26:23.000000000'
 '2019-07-28T15:27:07.000000000' '2019-07-28T15:57:59.000000000'
 '2019-07-28T16:30:04.000000000' '2019-07-28T19:46:20.000000000'
 '2019-07-29T05:09:26.000000000' '2019-07-30T19:04:23.000000000'
 '2019-07-30T22:33:45.000000000' '2019-08-12T19:49:14.000000000'
 '2019-08-12T22:06:25.000000000' '2019-08-12T23:23:18.000000000'
 '2019-08-13T01:17:21.000000000' '2019-08-13T03:04:56.000000000'
 '2019-08-13T06:10:36.000000000' '2019-08-13T09:26:22.000000000'
 '2019-08-13T10:30:20.000000000' '2019-08-13T12:50:54.000000000'
 '2019-08-13T13:33:02.000000000' '2019-08-13T15:13:20.000000000'
 '2019-08-13T16:09:14.000000000' '2019-08-13T16:48:03.000000000'
 '2019-08-13T18:19:24.000000000' '2019-08-13T20:59:00.000000000'
 '2019-08-13T22:23:13.000000000' '2019-08-13T22:51:40.000000000'
 '2019-08-14T03:53:23.000000000' '2019-08-14T04:41:52.000000000'
 '2019-08-18T12:23:50.000000000' '2019-08-18T13:32:48.000000000'
 '2019-08-18T14:17:09.000000000' '2019-08-18T14:51:35.000000000'
 '2019-08-18T15:43:21.000000000' '2019-08-18T16:20:32.000000000'
 '2019-08-18T17:05:42.000000000' '2019-08-18T18:25:03.000000000'
 '2019-08-18T19:10:56.000000000'                           'NaT']
Количество уникальных значений: 74
****************************************************************************************************
Уникальные значения столбца 'Wire 7':
['2019-07-27T05:49:05.000000000' '2019-07-27T07:56:34.000000000'
 '2019-07-27T13:43:32.000000000' '2019-07-27T19:11:22.000000000'
 '2019-08-12T19:47:06.000000000' '2019-08-12T21:48:11.000000000'
 '2019-08-12T23:20:37.000000000' '2019-08-13T01:13:45.000000000'
 '2019-08-13T02:52:06.000000000' '2019-08-13T07:49:49.000000000'
 '2019-08-13T10:25:22.000000000'                           'NaT']
Количество уникальных значений: 12
****************************************************************************************************
Уникальные значения столбца 'Wire 8':
['2019-05-14T11:29:24.000000000' '2019-05-14T12:18:01.000000000'
 '2019-05-14T12:52:37.000000000' '2019-05-14T13:20:41.000000000'
 '2019-05-14T14:05:13.000000000' '2019-05-14T14:35:02.000000000'
 '2019-05-14T15:07:59.000000000' '2019-05-14T15:43:01.000000000'
 '2019-05-14T16:18:12.000000000' '2019-05-14T16:55:09.000000000'
 '2019-07-08T16:56:51.000000000' '2019-07-22T08:33:41.000000000'
 '2019-07-22T10:05:30.000000000' '2019-08-16T04:37:28.000000000'
 '2019-08-16T05:36:19.000000000' '2019-08-16T06:42:52.000000000'
 '2019-08-16T07:39:37.000000000' '2019-08-16T08:14:32.000000000'
 '2019-08-16T08:56:23.000000000'                           'NaT']
Количество уникальных значений: 20
****************************************************************************************************
Уникальные значения столбца 'Wire 9':
['2019-05-04T17:21:27.000000000' '2019-06-02T15:42:11.000000000'
 '2019-06-02T18:02:35.000000000' '2019-06-02T18:51:06.000000000'
 '2019-06-02T20:35:26.000000000' '2019-06-02T23:08:17.000000000'
 '2019-06-09T11:01:10.000000000' '2019-06-09T19:36:17.000000000'
 '2019-06-09T19:55:31.000000000' '2019-06-09T20:29:31.000000000'
 '2019-06-09T21:38:20.000000000' '2019-06-09T22:22:10.000000000'
 '2019-06-09T22:50:56.000000000' '2019-06-11T07:58:12.000000000'
 '2019-06-11T10:10:21.000000000' '2019-07-04T21:52:41.000000000'
 '2019-07-04T23:02:08.000000000' '2019-07-05T00:44:55.000000000'
 '2019-07-05T05:14:22.000000000' '2019-07-05T05:40:57.000000000'
 '2019-08-09T06:04:40.000000000' '2019-08-09T06:49:04.000000000'
 '2019-08-09T08:01:55.000000000' '2019-08-09T09:11:30.000000000'
 '2019-08-09T10:17:13.000000000' '2019-08-09T11:16:48.000000000'
 '2019-08-09T13:07:31.000000000' '2019-08-09T14:03:42.000000000'
 '2019-09-03T12:55:23.000000000'                           'NaT']
Количество уникальных значений: 30
****************************************************************************************************


df_wire_time


df_wire_time['start_time_wire'] = df_wire_time.min(axis=1)
df_wire_time['finish_time_wire'] = df_wire_time.max(axis=1)
df_wire_time['duration_wire'] = (
    df_wire_time['finish_time_wire'] - df_wire_time['start_time_wire']).astype('timedelta64[s]')


df_wire_time


df_wire_time = df_wire_time[['start_time_wire', 'finish_time_wire','duration_wire']]
df_wire_time


# Сводная статистика для всех столбцов в `df_wire_time`
df_wire_time.describe(include='all', datetime_is_numeric=True)


# Посмотрим уникальные значения по колонкам `df_wire_time`
get_unique_sorted(df_wire_time)

Уникальные значения столбца 'start_time_wire':
['2019-05-03T11:06:19.000000000' '2019-05-03T11:36:50.000000000'
 '2019-05-03T12:11:46.000000000' ... '2019-09-06T14:36:11.000000000'
 '2019-09-06T15:33:55.000000000' '2019-09-06T17:10:06.000000000']
Количество уникальных значений: 3081
****************************************************************************************************
Уникальные значения столбца 'finish_time_wire':
['2019-05-03T11:06:19.000000000' '2019-05-03T11:36:50.000000000'
 '2019-05-03T12:11:46.000000000' ... '2019-09-06T14:36:11.000000000'
 '2019-09-06T15:33:55.000000000' '2019-09-06T17:10:06.000000000']
Количество уникальных значений: 3081
****************************************************************************************************
Уникальные значения столбца 'duration_wire':
[   0.   90.  125.  130.  132.  137.  141.  142.  143.  150.  151.  154.
  155.  157.  162.  165.  168.  171.  174.  176.  178.  179.  180.  183.
  184.  187.  193.  194.  196.  199.  200.  204.  207.  208.  209.  210.
  212.  217.  219.  220.  222.  224.  225.  226.  230.  231.  234.  235.
  236.  240.  241.  242.  243.  244.  245.  248.  249.  251.  252.  253.
  254.  257.  258.  259.  260.  261.  262.  263.  264.  265.  268.  269.
  271.  273.  274.  275.  276.  277.  280.  283.  284.  285.  286.  287.
  288.  290.  292.  293.  294.  295.  298.  299.  300.  301.  304.  305.
  306.  308.  309.  310.  311.  312.  313.  314.  315.  316.  317.  318.
  319.  320.  321.  322.  323.  324.  325.  326.  327.  329.  330.  331.
  332.  333.  334.  335.  336.  337.  338.  339.  340.  341.  342.  343.
  344.  345.  347.  348.  349.  350.  351.  352.  354.  355.  356.  357.
  358.  359.  360.  361.  362.  363.  364.  367.  369.  370.  371.  372.
  373.  374.  375.  376.  377.  378.  379.  380.  381.  382.  383.  384.
  385.  386.  387.  389.  390.  392.  393.  394.  395.  396.  397.  398.
  399.  400.  401.  402.  403.  404.  405.  406.  407.  408.  409.  410.
  411.  412.  413.  414.  415.  416.  417.  418.  419.  420.  421.  422.
  423.  424.  425.  426.  427.  428.  429.  430.  431.  432.  433.  434.
  435.  436.  437.  438.  439.  440.  441.  442.  443.  444.  445.  447.
  448.  449.  450.  451.  452.  453.  454.  455.  456.  457.  458.  459.
  460.  461.  462.  463.  464.  465.  466.  467.  468.  469.  470.  471.
  472.  473.  475.  476.  477.  478.  479.  480.  481.  482.  483.  484.
  485.  486.  487.  488.  489.  490.  491.  492.  493.  494.  496.  498.
  499.  502.  503.  504.  505.  506.  507.  508.  509.  510.  511.  513.
  514.  515.  516.  517.  518.  520.  521.  523.  525.  526.  527.  528.
  529.  530.  531.  532.  533.  535.  536.  537.  538.  540.  541.  542.
  543.  544.  545.  546.  547.  548.  549.  550.  551.  552.  553.  554.
  555.  556.  558.  561.  563.  564.  565.  566.  568.  571.  573.  574.
  575.  576.  578.  579.  580.  583.  584.  585.  586.  587.  588.  589.
  590.  593.  594.  595.  596.  597.  598.  599.  600.  601.  602.  603.
  604.  605.  606.  607.  609.  610.  611.  612.  615.  616.  618.  619.
  620.  621.  622.  623.  624.  625.  626.  627.  628.  629.  630.  631.
  632.  633.  634.  636.  638.  639.  644.  645.  646.  647.  648.  649.
  651.  652.  653.  654.  657.  659.  660.  663.  665.  666.  667.  668.
  669.  670.  672.  673.  674.  676.  678.  682.  683.  684.  685.  688.
  690.  691.  694.  695.  697.  698.  699.  700.  704.  705.  706.  708.
  711.  715.  718.  719.  722.  723.  726.  727.  729.  730.  731.  733.
  734.  735.  736.  737.  738.  739.  740.  741.  744.  747.  748.  749.
  750.  751.  754.  755.  756.  757.  759.  760.  761.  762.  764.  765.
  767.  768.  771.  774.  775.  776.  777.  780.  781.  783.  784.  785.
  786.  787.  790.  791.  792.  794.  797.  800.  801.  802.  803.  807.
  808.  812.  813.  817.  819.  830.  835.  839.  841.  842.  847.  850.
  853.  854.  856.  859.  860.  861.  865.  867.  868.  871.  875.  876.
  879.  881.  889.  894.  900.  902.  904.  915.  919.  921.  923.  925.
  926.  932.  937.  943.  957.  961.  973.  987.  994.  995. 1004. 1005.
 1007. 1015. 1019. 1030. 1034. 1036. 1037. 1039. 1047. 1048. 1050. 1051.
 1052. 1062. 1072. 1075. 1077. 1080. 1082. 1093. 1123. 1135. 1140. 1144.
 1154. 1180. 1183. 1222. 1224. 1236. 1277. 1324. 1329. 1333. 1342. 1364.
 1369. 1383. 1388. 1421. 1440. 1454. 1495. 1496. 1504. 1548. 1558. 1559.
 1580. 1600. 1615. 1664. 1670. 1698. 1704. 1728. 1777. 1799. 1861. 1875.
 1888. 1902. 1926. 1963. 1975. 2000. 2003. 2154. 2162. 2197. 2212. 2241.
 2252. 2351. 2535. 2548. 2761. 2765. 3375. 3455. 3477. 3578. 5013. 5492.
 5937.]
Количество уникальных значений: 625
****************************************************************************************************


# Создаем графическую фигуру
plt.figure(figsize=(15, 8))
# Pазбиваем графическую фигуру на 2 графика
# Первый график
plt.subplot(2, 1, 1)
# Гистограмма для 'duration_wire'
sns.histplot(data=df_wire_time[df_wire_time['duration_wire'] != 0]['duration_wire'],
             alpha=0.8,
             label='Длительность обработки с применением проволочных материалов (с)')

plt.minorticks_on()
plt.grid(which='minor',
         linestyle=':')
plt.grid(True)
plt.xlabel('')
plt.ylabel('Частота',
           fontsize=12,
           color='DarkSlateGray')
plt.legend()

# Второй график
plt.subplot(2, 1, 2)
# Боксплот для 'duration_wire'
sns.boxplot(data=df_wire_time[df_wire_time['duration_wire'] != 0],
            x='duration_wire',
            orient='horizontal')

plt.minorticks_on()
plt.grid(which='minor',
         linestyle=':')
plt.grid(True)
plt.xlabel('Длительность обработки с применением проволочных материалов (с)',
           fontsize=12,
           color='DarkSlateGray')
plt.ylabel('')
plt.suptitle('Распределение длительности обработки с применением проволочных материалов (с)',
             fontsize=15,
             color='DarkSlateGray')
plt.show()


# Удалим редкий сыпучий материал
df_bulk = df_bulk.drop('bulk_8', axis=1)
# Удалим данные с температурами ниже температуры плавления
df_result_temp = df_result_temp.drop(df_result_temp[(df_result_temp['temperature_first_measurement'] < 1300) | (
    df_result_temp['temperature_last_measurement'] < 1300)].index)
# Удалим редкий проволочный материал
df_wire = df_wire.drop('wire_5', axis=1)


display(df_arc_key, df_bulk, df_bulk_time, df_gas, df_result_temp, df_wire, df_wire_time)


# Объединим 7 таблиц в одну (присоединять будем к самой короткой)
unified_table = df_result_temp.join(
    [df_arc_key, df_bulk, df_bulk_time, df_gas, df_wire, df_wire_time], how='inner')
unified_table


''' Подсчитываем количество строк в `unified_table`, 
где столбец `time_last_measurement` не является максимальным значением среди
всех столбцов с типом данных `datetime` в `unified_table`.
'''
unified_table[
    (
        unified_table['time_last_measurement'] >= unified_table[unified_table.select_dtypes(
            'datetime').columns].max(axis=1)
    )
    .astype(int) == 0][unified_table.select_dtypes('datetime').columns].shape[0]

0


# удалим столбцы с типом данных `datetime`
unified_table = unified_table.drop(unified_table.select_dtypes('datetime').columns, axis=1)
unified_table


# Функция визуализации сильно коррелирующих переменных
def plot_highly_correlated_features(df):
    # Вычисляем матрицу корреляции
    high_corr_pairs = df.corr().unstack()
    # Оставляем только пары, где корреляция от 0.7 до 1
    high_corr_pairs = high_corr_pairs[(high_corr_pairs < 1) & (abs(high_corr_pairs) > 0.7)]
    # Проверяем, есть ли сильно коррелирующие признаки
    if high_corr_pairs.empty:
        print("Сильно коррелирующих признаков нет")
    else:
        # Извлекаем уникальные имена переменных из индекса полученных пар
        high_corr_features = high_corr_pairs.index.get_level_values(0).unique()
        # Создаем новый DataFrame, содержащий только сильно коррелирующие переменные
        highly_corr_df = df[high_corr_features]
        # Вычисляем корреляционную матрицу для сильно коррелирующих переменных
        corr_matrix = highly_corr_df.corr()
        # Устанавливаем размер и цвет графика
        sns.set(rc={'figure.figsize': (15, 5)})
        # Создаем тепловую карту корреляционной матрицы
        sns.heatmap(corr_matrix,
                    annot=True,
                    cmap='turbo',
                    vmin=-1,
                    vmax=1,
                    center=0,
                    cbar_kws={'orientation': 'vertical'},
                    mask=np.triu(np.ones_like(corr_matrix)))
        # Добавляем название графику
        plt.title('Корреляционная матрица сильно коррелирующих переменных',
                  fontsize=15,
                  color='DarkSlateGray')
        plt.show()


plot_highly_correlated_features(unified_table)


# Удалим столбцы с активной, реактивной и полной мощностью, с продолжительностью дуги
unified_table.drop(['active_power', 'reactive_power',
                   'arc_duration', 'apparent_power'], axis=1, inplace=True)


plot_highly_correlated_features(unified_table)


# Удалим столбцы с количеством этапов обработки и общим количеством подаваемых сыпучих материалов.
unified_table.drop(['count_arc', 'bulk_sum'], axis=1, inplace=True)
# Добавление столбца bulk_7_wire_4
unified_table['bulk_7_wire_4'] = unified_table['bulk_7'] + unified_table['wire_4']
# Добавление столбца bulk_9_wire_8
unified_table['bulk_9_wire_8'] = unified_table['bulk_9'] + unified_table['wire_8']
# Удаление столбцов bulk_7, wire_4, bulk_9, wire_8
unified_table.drop(['bulk_7', 'wire_4', 'bulk_9', 'wire_8'], axis=1, inplace=True)


unified_table


plot_highly_correlated_features(unified_table)


# Добавление столбца bulk_7_wire_4_bulk_2
unified_table['bulk_7_wire_4_bulk_2'] = unified_table['bulk_7_wire_4'] + unified_table['bulk_2']
# Удаление столбцов bulk_7_wire_4, bulk_2
unified_table.drop(['bulk_7_wire_4', 'bulk_2'], axis=1, inplace=True)
plot_highly_correlated_features(unified_table)

Сильно коррелирующих признаков нет


unified_table


# Создаем пустую серию для хранения корреляций с целевым столбцом
corr_target = pd.Series(dtype='float64')
# Итерируемся по столбцам таблицы
for col in unified_table.columns:
    # Считаем корреляцию между столбцом и целевым столбцом и добавляем в серию
    corr_target[col] = unified_table['temperature_last_measurement'].corr(
        unified_table[col])
# Сортируем серию по убыванию и исключаем целевой столбец
corr_target = corr_target.sort_values(ascending=False)[1:]
# Создаем график
plt.figure(figsize=(15, 5))
sns.barplot(x=corr_target.index,
            y=corr_target.values)
plt.xticks(rotation=90)
plt.title('Корреляция признаков с последним измерением температуры',
          fontsize=15,
          color='DarkSlateGray')
plt.xlabel('Признаки',
           fontsize=12,
           color='DarkSlateGray')
plt.ylabel('Корреляция',
           fontsize=12,
           color='DarkSlateGray')
plt.minorticks_on()
plt.grid(which='minor',
         linestyle=':')
plt.grid(True)
plt.show()


# Разделим данные на обучающую и тестовую выборки
train, test = train_test_split(
    unified_table, test_size=0.25, random_state=RANDOM_STATE)
# Определим целевой признак для обучающей выборки
target_train = train['temperature_last_measurement']
# Определим признаки для обучающей выборки
features_train = train.drop(['temperature_last_measurement'], axis=1)
# Определим целевой признак для тестовой выборки
target_test = test['temperature_last_measurement']
# Определим признаки для тестовой выборки
features_test = test.drop(['temperature_last_measurement'], axis=1)
# Отобразим результат
display(target_train, features_train, target_test, features_test)

key
1973    1579.0
802     1592.0
959     1617.0
680     1588.0
970     1579.0
         ...  
119     1593.0
809     1604.0
1775    1596.0
1966    1578.0
1580    1593.0
Name: temperature_last_measurement, Length: 1743, dtype: float64

key
1480    1597.0
2322    1590.0
673     1614.0
660     1590.0
1357    1600.0
         ...  
335     1612.0
2151    1620.0
537     1585.0
650     1621.0
569     1625.0
Name: temperature_last_measurement, Length: 582, dtype: float64


# Получаем названия числовых признаков
numeric = features_train.columns
# Создаем экземпляр стандартизатора
scaler = StandardScaler()
# Обучаем и проводим стандартизацию на тренировочных данных
scaler.fit(features_train[numeric])
features_train[numeric] = scaler.transform(features_train[numeric])
# Проводим ту же стандартизацию на тестовых данных
features_test[numeric] = scaler.transform(features_test[numeric])
# Выводим преобразованные тренировочные и тестовые данные
display(features_train, features_test)


round(mae(target_train, np.ones(features_train.shape[0])*target_train.mean()), 2)

7.88


# Cоздадим объект `KFold`, который будем использовать для разделения данных
kf = KFold(n_splits=5, shuffle=True, random_state=RANDOM_STATE)


# Создаем Pipeline с масштабированием и регрессией
pipeline = make_pipeline(StandardScaler(), regressor)
# Вычисляем значение метрики с помощью кросс-валидации
cv_MAE_LR = round(abs(cross_val_score(pipeline,
                                      features_train,
                                      target_train,
                                      cv=kf,
                                      scoring='neg_mean_absolute_error', n_jobs=-1)).mean(), 2)
print('Значение метрики:', cv_MAE_LR)

Значение метрики: 6.55
Wall time: 2.68 s


%%time

def objective(trial):
    # Определяем параметры для подбора
    n_estimators = trial.suggest_int('n_estimators', 10, 500)
    max_depth = trial.suggest_int('max_depth', 5, 30)
    min_samples_split = trial.suggest_int('min_samples_split', 2, 6)
    min_samples_leaf = trial.suggest_int('min_samples_leaf', 1, 3)
    max_features = trial.suggest_categorical(
        'max_features', ['sqrt', 'log2', None])
    bootstrap = trial.suggest_categorical('bootstrap', [True, False])

# Создадём модель случайного леса с определенными параметрами
    model = RandomForestRegressor(n_estimators=n_estimators,
                                  max_depth=max_depth,
                                  min_samples_split=min_samples_split,
                                  min_samples_leaf=min_samples_leaf,
                                  max_features=max_features,
                                  bootstrap=bootstrap,
                                  random_state=RANDOM_STATE,
                                  n_jobs=-1)

# Оцениваем качество модели с помощью кросс-валидации
    score = round(abs(cross_val_score(model,
                                      features_train,
                                      target_train,
                                      cv=kf,
                                      scoring='neg_mean_absolute_error', n_jobs=-1)).mean(), 2)
# Возврщаем значения метрики
    return score.mean()


# Создаём объект Study и запускаем оптимизацию
study = optuna.create_study(direction='minimize')
study.optimize(objective, n_trials=50)

print('Лучшие параметры:', study.best_params)
print('Лучшее значение метрики:', study.best_value)

# Создаём датафрейм с значениями гиперпараметров и метрикой
results = pd.DataFrame({'n_estimators': [],
                       'max_depth': [],
                        'min_samples_split': [],
                        'min_samples_leaf': [],
                        'max_features': [],
                        'bootstrap': [],
                        'score': []})
# Заполняем данными из best_trials
for trial in study.best_trials:
    params = trial.params
    params['score'] = trial.value
    results = results.append(params, ignore_index=True)
# Сортируем датафрейм по убыванию значения метрики
results = results.sort_values(by='score', ascending=False)
# Выводим 10 лучших моделей
display(results.head(10))

# Создаём график важности параметров
optuna.visualization.matplotlib.plot_param_importances(study)
plt.xlabel('Важность гиперпараметров',
           fontsize=12,
           color='DarkSlateGray')
plt.ylabel('Гиперпараметры',
           fontsize=12,
           color='DarkSlateGray')
plt.suptitle('Важность гиперпараметров',
             fontsize=15,
             color='DarkSlateGray')
plt.minorticks_on()
plt.grid(which='minor',
         linestyle=':')
plt.grid(True)

#  Создаём график истории оптимизации
optuna.visualization.matplotlib.plot_optimization_history(study)
plt.tight_layout()
plt.xlabel('Итерации',
           fontsize=12,
           color='DarkSlateGray')
plt.ylabel('Целевая метрика',
           fontsize=12,
           color='DarkSlateGray')
plt.suptitle('История оптимизации \n \n',
             fontsize=15,
             color='DarkSlateGray')
plt.minorticks_on()
plt.grid(which='minor',
         linestyle=':')
plt.grid(True)
plt.show()

[I 2023-08-24 10:15:19,986] A new study created in memory with name: no-name-c5093fcd-01c0-457c-9ffc-6126a90c4d3a
[I 2023-08-24 10:15:27,559] Trial 0 finished with value: 8.16 and parameters: {'n_estimators': 476, 'max_depth': 15, 'min_samples_split': 4, 'min_samples_leaf': 1, 'max_features': None, 'bootstrap': False}. Best is trial 0 with value: 8.16.
[I 2023-08-24 10:15:29,151] Trial 1 finished with value: 6.65 and parameters: {'n_estimators': 428, 'max_depth': 14, 'min_samples_split': 2, 'min_samples_leaf': 1, 'max_features': 'log2', 'bootstrap': True}. Best is trial 1 with value: 6.65.
[I 2023-08-24 10:15:30,342] Trial 2 finished with value: 6.62 and parameters: {'n_estimators': 343, 'max_depth': 22, 'min_samples_split': 4, 'min_samples_leaf': 3, 'max_features': 'sqrt', 'bootstrap': True}. Best is trial 2 with value: 6.62.
[I 2023-08-24 10:15:30,761] Trial 3 finished with value: 6.7 and parameters: {'n_estimators': 123, 'max_depth': 28, 'min_samples_split': 5, 'min_samples_leaf': 3, 'max_features': 'log2', 'bootstrap': True}. Best is trial 2 with value: 6.62.
[I 2023-08-24 10:15:31,851] Trial 4 finished with value: 6.59 and parameters: {'n_estimators': 259, 'max_depth': 28, 'min_samples_split': 4, 'min_samples_leaf': 2, 'max_features': 'log2', 'bootstrap': False}. Best is trial 4 with value: 6.59.
[I 2023-08-24 10:15:34,369] Trial 5 finished with value: 6.59 and parameters: {'n_estimators': 473, 'max_depth': 30, 'min_samples_split': 4, 'min_samples_leaf': 1, 'max_features': 'sqrt', 'bootstrap': False}. Best is trial 4 with value: 6.59.
[I 2023-08-24 10:15:34,568] Trial 6 finished with value: 6.78 and parameters: {'n_estimators': 31, 'max_depth': 30, 'min_samples_split': 2, 'min_samples_leaf': 1, 'max_features': 'log2', 'bootstrap': False}. Best is trial 4 with value: 6.59.
[I 2023-08-24 10:15:39,684] Trial 7 finished with value: 6.45 and parameters: {'n_estimators': 451, 'max_depth': 29, 'min_samples_split': 3, 'min_samples_leaf': 1, 'max_features': None, 'bootstrap': True}. Best is trial 7 with value: 6.45.
[I 2023-08-24 10:15:40,707] Trial 8 finished with value: 6.77 and parameters: {'n_estimators': 362, 'max_depth': 8, 'min_samples_split': 4, 'min_samples_leaf': 2, 'max_features': 'log2', 'bootstrap': True}. Best is trial 7 with value: 6.45.
[I 2023-08-24 10:15:41,719] Trial 9 finished with value: 6.67 and parameters: {'n_estimators': 301, 'max_depth': 13, 'min_samples_split': 3, 'min_samples_leaf': 1, 'max_features': 'log2', 'bootstrap': True}. Best is trial 7 with value: 6.45.
[I 2023-08-24 10:15:43,612] Trial 10 finished with value: 6.44 and parameters: {'n_estimators': 179, 'max_depth': 22, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 10 with value: 6.44.
[I 2023-08-24 10:15:45,237] Trial 11 finished with value: 6.44 and parameters: {'n_estimators': 160, 'max_depth': 23, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 10 with value: 6.44.
[I 2023-08-24 10:15:47,012] Trial 12 finished with value: 6.45 and parameters: {'n_estimators': 172, 'max_depth': 22, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 10 with value: 6.44.
[I 2023-08-24 10:15:48,750] Trial 13 finished with value: 6.45 and parameters: {'n_estimators': 171, 'max_depth': 21, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 10 with value: 6.44.
[I 2023-08-24 10:15:49,611] Trial 14 finished with value: 6.44 and parameters: {'n_estimators': 88, 'max_depth': 19, 'min_samples_split': 6, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 10 with value: 6.44.
[I 2023-08-24 10:15:51,790] Trial 15 finished with value: 6.45 and parameters: {'n_estimators': 211, 'max_depth': 25, 'min_samples_split': 5, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 10 with value: 6.44.
[I 2023-08-24 10:15:52,044] Trial 16 finished with value: 6.61 and parameters: {'n_estimators': 20, 'max_depth': 24, 'min_samples_split': 5, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 10 with value: 6.44.
[I 2023-08-24 10:15:53,041] Trial 17 finished with value: 6.43 and parameters: {'n_estimators': 102, 'max_depth': 18, 'min_samples_split': 6, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 17 with value: 6.43.
[I 2023-08-24 10:15:53,440] Trial 18 finished with value: 6.62 and parameters: {'n_estimators': 83, 'max_depth': 17, 'min_samples_split': 5, 'min_samples_leaf': 3, 'max_features': 'sqrt', 'bootstrap': False}. Best is trial 17 with value: 6.43.
[I 2023-08-24 10:15:55,389] Trial 19 finished with value: 6.45 and parameters: {'n_estimators': 245, 'max_depth': 9, 'min_samples_split': 6, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 17 with value: 6.43.
[I 2023-08-24 10:15:56,308] Trial 20 finished with value: 6.43 and parameters: {'n_estimators': 102, 'max_depth': 11, 'min_samples_split': 5, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 17 with value: 6.43.
[I 2023-08-24 10:15:56,847] Trial 21 finished with value: 6.65 and parameters: {'n_estimators': 89, 'max_depth': 5, 'min_samples_split': 5, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 17 with value: 6.43.
[I 2023-08-24 10:15:57,895] Trial 22 finished with value: 6.44 and parameters: {'n_estimators': 116, 'max_depth': 11, 'min_samples_split': 6, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 17 with value: 6.43.
[I 2023-08-24 10:15:59,886] Trial 23 finished with value: 6.44 and parameters: {'n_estimators': 205, 'max_depth': 19, 'min_samples_split': 5, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 17 with value: 6.43.
[I 2023-08-24 10:16:00,380] Trial 24 finished with value: 6.47 and parameters: {'n_estimators': 46, 'max_depth': 17, 'min_samples_split': 6, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 17 with value: 6.43.
[I 2023-08-24 10:16:01,041] Trial 25 finished with value: 6.65 and parameters: {'n_estimators': 143, 'max_depth': 19, 'min_samples_split': 5, 'min_samples_leaf': 2, 'max_features': 'sqrt', 'bootstrap': True}. Best is trial 17 with value: 6.43.
[I 2023-08-24 10:16:01,880] Trial 26 finished with value: 8.08 and parameters: {'n_estimators': 61, 'max_depth': 12, 'min_samples_split': 6, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': False}. Best is trial 17 with value: 6.43.
[I 2023-08-24 10:16:03,873] Trial 27 finished with value: 6.45 and parameters: {'n_estimators': 198, 'max_depth': 16, 'min_samples_split': 5, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 17 with value: 6.43.
[I 2023-08-24 10:16:05,138] Trial 28 finished with value: 6.44 and parameters: {'n_estimators': 118, 'max_depth': 25, 'min_samples_split': 6, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 17 with value: 6.43.
[I 2023-08-24 10:16:07,139] Trial 29 finished with value: 7.21 and parameters: {'n_estimators': 236, 'max_depth': 6, 'min_samples_split': 5, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': False}. Best is trial 17 with value: 6.43.
[I 2023-08-24 10:16:07,413] Trial 30 finished with value: 6.7 and parameters: {'n_estimators': 65, 'max_depth': 15, 'min_samples_split': 6, 'min_samples_leaf': 3, 'max_features': 'sqrt', 'bootstrap': True}. Best is trial 17 with value: 6.43.
[I 2023-08-24 10:16:08,995] Trial 31 finished with value: 6.45 and parameters: {'n_estimators': 157, 'max_depth': 21, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 17 with value: 6.43.
[I 2023-08-24 10:16:10,292] Trial 32 finished with value: 6.44 and parameters: {'n_estimators': 126, 'max_depth': 23, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 17 with value: 6.43.
[I 2023-08-24 10:16:12,241] Trial 33 finished with value: 6.45 and parameters: {'n_estimators': 182, 'max_depth': 26, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 17 with value: 6.43.
[I 2023-08-24 10:16:13,819] Trial 34 finished with value: 6.45 and parameters: {'n_estimators': 144, 'max_depth': 20, 'min_samples_split': 5, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 17 with value: 6.43.
[I 2023-08-24 10:16:16,801] Trial 35 finished with value: 6.44 and parameters: {'n_estimators': 296, 'max_depth': 15, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 17 with value: 6.43.
[I 2023-08-24 10:16:17,867] Trial 36 finished with value: 6.43 and parameters: {'n_estimators': 107, 'max_depth': 27, 'min_samples_split': 5, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 17 with value: 6.43.
[I 2023-08-24 10:16:18,448] Trial 37 finished with value: 6.61 and parameters: {'n_estimators': 108, 'max_depth': 27, 'min_samples_split': 4, 'min_samples_leaf': 3, 'max_features': 'sqrt', 'bootstrap': False}. Best is trial 17 with value: 6.43.
[I 2023-08-24 10:16:18,693] Trial 38 finished with value: 6.74 and parameters: {'n_estimators': 57, 'max_depth': 11, 'min_samples_split': 3, 'min_samples_leaf': 3, 'max_features': 'log2', 'bootstrap': True}. Best is trial 17 with value: 6.43.
[I 2023-08-24 10:16:19,647] Trial 39 finished with value: 6.45 and parameters: {'n_estimators': 86, 'max_depth': 28, 'min_samples_split': 5, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 17 with value: 6.43.
[I 2023-08-24 10:16:19,785] Trial 40 finished with value: 6.74 and parameters: {'n_estimators': 11, 'max_depth': 13, 'min_samples_split': 4, 'min_samples_leaf': 3, 'max_features': 'log2', 'bootstrap': False}. Best is trial 17 with value: 6.43.
[I 2023-08-24 10:16:21,464] Trial 41 finished with value: 6.45 and parameters: {'n_estimators': 131, 'max_depth': 24, 'min_samples_split': 5, 'min_samples_leaf': 1, 'max_features': None, 'bootstrap': True}. Best is trial 17 with value: 6.43.
[I 2023-08-24 10:16:23,794] Trial 42 finished with value: 6.44 and parameters: {'n_estimators': 225, 'max_depth': 22, 'min_samples_split': 6, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 17 with value: 6.43.
[I 2023-08-24 10:16:26,847] Trial 43 finished with value: 6.44 and parameters: {'n_estimators': 271, 'max_depth': 26, 'min_samples_split': 4, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 17 with value: 6.43.
[I 2023-08-24 10:16:28,553] Trial 44 finished with value: 6.44 and parameters: {'n_estimators': 156, 'max_depth': 20, 'min_samples_split': 3, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 17 with value: 6.43.
[I 2023-08-24 10:16:30,628] Trial 45 finished with value: 6.46 and parameters: {'n_estimators': 185, 'max_depth': 23, 'min_samples_split': 6, 'min_samples_leaf': 1, 'max_features': None, 'bootstrap': True}. Best is trial 17 with value: 6.43.
[I 2023-08-24 10:16:31,083] Trial 46 finished with value: 6.74 and parameters: {'n_estimators': 104, 'max_depth': 18, 'min_samples_split': 2, 'min_samples_leaf': 3, 'max_features': 'log2', 'bootstrap': True}. Best is trial 17 with value: 6.43.
[I 2023-08-24 10:16:35,495] Trial 47 finished with value: 6.44 and parameters: {'n_estimators': 422, 'max_depth': 30, 'min_samples_split': 5, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 17 with value: 6.43.
[I 2023-08-24 10:16:36,133] Trial 48 finished with value: 6.65 and parameters: {'n_estimators': 168, 'max_depth': 21, 'min_samples_split': 6, 'min_samples_leaf': 3, 'max_features': 'sqrt', 'bootstrap': True}. Best is trial 17 with value: 6.43.
[I 2023-08-24 10:16:36,534] Trial 49 finished with value: 6.57 and parameters: {'n_estimators': 32, 'max_depth': 28, 'min_samples_split': 5, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 17 with value: 6.43.

Лучшие параметры: {'n_estimators': 102, 'max_depth': 18, 'min_samples_split': 6, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}
Лучшее значение метрики: 6.43

<timed exec>:57: ExperimentalWarning:

plot_param_importances is experimental (supported from v2.2.0). The interface can change in the future.

<timed exec>:73: ExperimentalWarning:

plot_optimization_history is experimental (supported from v2.2.0). The interface can change in the future.

Wall time: 1min 24s


# Создаём график контура
fig = optuna.visualization.plot_contour(
    study, params=['max_depth', 'n_estimators'])
fig.show()


%%time

def objective(trial):
    # Определяем параметры для подбора
    n_estimators = trial.suggest_int('n_estimators', 14, 450)
    max_depth = trial.suggest_int('max_depth', 10, 26)
    min_samples_split = trial.suggest_int('min_samples_split', 3, 6)
    min_samples_leaf = trial.suggest_int('min_samples_leaf', 1, 3)
    max_features = trial.suggest_categorical('max_features', [None])
    bootstrap = trial.suggest_categorical('bootstrap', [True])

# Создадём модель случайного леса с определенными параметрами
    model = RandomForestRegressor(n_estimators=n_estimators,
                                  max_depth=max_depth,
                                  min_samples_split=min_samples_split,
                                  min_samples_leaf=min_samples_leaf,
                                  max_features=max_features,
                                  bootstrap=bootstrap,
                                  random_state=RANDOM_STATE, 
                                  n_jobs=-1)

# Оцениваем качество модели с помощью кросс-валидации
    score = round(abs(cross_val_score(model,
                                      features_train,
                                      target_train,
                                      cv=kf,
                                      scoring='neg_mean_absolute_error', n_jobs=-1)).mean(), 2)
# Возврщаем значения метрики
    return score.mean()


# Создаём объект Study и запускаем оптимизацию
study = optuna.create_study(direction='minimize')
study.optimize(objective, n_trials=50)

print('Лучшие параметры:', study.best_params)
print('Лучшее значение метрики:', study.best_value)

# Создаём датафрейм с значениями гиперпараметров и метрикой
results = pd.DataFrame({'n_estimators': [],
                       'max_depth': [],
                        'min_samples_split': [],
                        'min_samples_leaf': [],
                        'max_features': [],
                        'bootstrap': [],
                        'score': []})
# Заполняем данными из best_trials
for trial in study.best_trials:
    params = trial.params
    params['score'] = trial.value
    results = results.append(params, ignore_index=True)
# Сортируем датафрейм по убыванию значения метрики
results = results.sort_values(by='score', ascending=False)
# Выводим 10 лучших моделей
display(results.head(10))

# Создаём график важности параметров
optuna.visualization.matplotlib.plot_param_importances(study)
plt.xlabel('Важность гиперпараметров',
           fontsize=12,
           color='DarkSlateGray')
plt.ylabel('Гиперпараметры',
           fontsize=12,
           color='DarkSlateGray')
plt.suptitle('Важность гиперпараметров',
             fontsize=15,
             color='DarkSlateGray')
plt.minorticks_on()
plt.grid(which='minor',
         linestyle=':')
plt.grid(True)

#  Создаём график истории оптимизации
optuna.visualization.matplotlib.plot_optimization_history(study)
plt.tight_layout()
plt.xlabel('Итерации',
           fontsize=12,
           color='DarkSlateGray')
plt.ylabel('Целевая метрика',
           fontsize=12,
           color='DarkSlateGray')
plt.suptitle('История оптимизации \n \n',
             fontsize=15,
             color='DarkSlateGray')
plt.minorticks_on()
plt.grid(which='minor',
         linestyle=':')
plt.grid(True)
plt.show()

[I 2023-08-24 10:16:44,629] A new study created in memory with name: no-name-e5ecc5e4-5c56-477e-b241-b6f1db4aba45
[I 2023-08-24 10:16:47,215] Trial 0 finished with value: 6.46 and parameters: {'n_estimators': 211, 'max_depth': 25, 'min_samples_split': 4, 'min_samples_leaf': 1, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.46.
[I 2023-08-24 10:16:50,149] Trial 1 finished with value: 6.45 and parameters: {'n_estimators': 262, 'max_depth': 16, 'min_samples_split': 5, 'min_samples_leaf': 1, 'max_features': None, 'bootstrap': True}. Best is trial 1 with value: 6.45.
[I 2023-08-24 10:16:50,958] Trial 2 finished with value: 6.44 and parameters: {'n_estimators': 84, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 2 with value: 6.44.
[I 2023-08-24 10:16:52,729] Trial 3 finished with value: 6.46 and parameters: {'n_estimators': 157, 'max_depth': 18, 'min_samples_split': 4, 'min_samples_leaf': 1, 'max_features': None, 'bootstrap': True}. Best is trial 2 with value: 6.44.
[I 2023-08-24 10:16:55,803] Trial 4 finished with value: 6.44 and parameters: {'n_estimators': 315, 'max_depth': 17, 'min_samples_split': 4, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 2 with value: 6.44.
[I 2023-08-24 10:16:59,289] Trial 5 finished with value: 6.45 and parameters: {'n_estimators': 355, 'max_depth': 23, 'min_samples_split': 3, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 2 with value: 6.44.
[I 2023-08-24 10:17:01,185] Trial 6 finished with value: 6.44 and parameters: {'n_estimators': 190, 'max_depth': 18, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 2 with value: 6.44.
[I 2023-08-24 10:17:02,535] Trial 7 finished with value: 6.45 and parameters: {'n_estimators': 148, 'max_depth': 13, 'min_samples_split': 6, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 2 with value: 6.44.
[I 2023-08-24 10:17:03,386] Trial 8 finished with value: 6.44 and parameters: {'n_estimators': 81, 'max_depth': 21, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 2 with value: 6.44.
[I 2023-08-24 10:17:04,882] Trial 9 finished with value: 6.44 and parameters: {'n_estimators': 157, 'max_depth': 18, 'min_samples_split': 3, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 2 with value: 6.44.
[I 2023-08-24 10:17:05,077] Trial 10 finished with value: 6.57 and parameters: {'n_estimators': 16, 'max_depth': 10, 'min_samples_split': 5, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 2 with value: 6.44.
[I 2023-08-24 10:17:09,008] Trial 11 finished with value: 6.44 and parameters: {'n_estimators': 431, 'max_depth': 14, 'min_samples_split': 4, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 2 with value: 6.44.
[I 2023-08-24 10:17:11,792] Trial 12 finished with value: 6.44 and parameters: {'n_estimators': 312, 'max_depth': 12, 'min_samples_split': 5, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 2 with value: 6.44.
[I 2023-08-24 10:17:15,241] Trial 13 finished with value: 6.44 and parameters: {'n_estimators': 363, 'max_depth': 15, 'min_samples_split': 4, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 2 with value: 6.44.
[I 2023-08-24 10:17:18,133] Trial 14 finished with value: 6.44 and parameters: {'n_estimators': 272, 'max_depth': 21, 'min_samples_split': 5, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 2 with value: 6.44.
[I 2023-08-24 10:17:18,904] Trial 15 finished with value: 6.44 and parameters: {'n_estimators': 80, 'max_depth': 10, 'min_samples_split': 3, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 2 with value: 6.44.
[I 2023-08-24 10:17:23,544] Trial 16 finished with value: 6.44 and parameters: {'n_estimators': 431, 'max_depth': 16, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 2 with value: 6.44.
[I 2023-08-24 10:17:23,784] Trial 17 finished with value: 6.58 and parameters: {'n_estimators': 15, 'max_depth': 21, 'min_samples_split': 5, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 2 with value: 6.44.
[I 2023-08-24 10:17:24,654] Trial 18 finished with value: 6.43 and parameters: {'n_estimators': 85, 'max_depth': 12, 'min_samples_split': 4, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 18 with value: 6.43.
[I 2023-08-24 10:17:25,707] Trial 19 finished with value: 6.44 and parameters: {'n_estimators': 98, 'max_depth': 12, 'min_samples_split': 3, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 18 with value: 6.43.
[I 2023-08-24 10:17:26,849] Trial 20 finished with value: 6.43 and parameters: {'n_estimators': 111, 'max_depth': 12, 'min_samples_split': 6, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 18 with value: 6.43.
[I 2023-08-24 10:17:27,868] Trial 21 finished with value: 6.43 and parameters: {'n_estimators': 106, 'max_depth': 11, 'min_samples_split': 6, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 18 with value: 6.43.
[I 2023-08-24 10:17:28,983] Trial 22 finished with value: 6.44 and parameters: {'n_estimators': 118, 'max_depth': 11, 'min_samples_split': 6, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 18 with value: 6.43.
[I 2023-08-24 10:17:29,514] Trial 23 finished with value: 6.46 and parameters: {'n_estimators': 53, 'max_depth': 13, 'min_samples_split': 5, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 18 with value: 6.43.
[I 2023-08-24 10:17:30,652] Trial 24 finished with value: 6.44 and parameters: {'n_estimators': 128, 'max_depth': 11, 'min_samples_split': 6, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 18 with value: 6.43.
[I 2023-08-24 10:17:31,211] Trial 25 finished with value: 6.43 and parameters: {'n_estimators': 54, 'max_depth': 12, 'min_samples_split': 5, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 18 with value: 6.43.
[I 2023-08-24 10:17:32,814] Trial 26 finished with value: 6.44 and parameters: {'n_estimators': 198, 'max_depth': 10, 'min_samples_split': 4, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 18 with value: 6.43.
[I 2023-08-24 10:17:33,330] Trial 27 finished with value: 6.45 and parameters: {'n_estimators': 50, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 18 with value: 6.43.
[I 2023-08-24 10:17:34,855] Trial 28 finished with value: 6.44 and parameters: {'n_estimators': 175, 'max_depth': 11, 'min_samples_split': 5, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 18 with value: 6.43.
[I 2023-08-24 10:17:37,167] Trial 29 finished with value: 6.45 and parameters: {'n_estimators': 233, 'max_depth': 13, 'min_samples_split': 4, 'min_samples_leaf': 1, 'max_features': None, 'bootstrap': True}. Best is trial 18 with value: 6.43.
[I 2023-08-24 10:17:38,364] Trial 30 finished with value: 6.45 and parameters: {'n_estimators': 122, 'max_depth': 15, 'min_samples_split': 6, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 18 with value: 6.43.
[I 2023-08-24 10:17:38,922] Trial 31 finished with value: 6.48 and parameters: {'n_estimators': 48, 'max_depth': 26, 'min_samples_split': 5, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 18 with value: 6.43.
[I 2023-08-24 10:17:39,448] Trial 32 finished with value: 6.43 and parameters: {'n_estimators': 54, 'max_depth': 12, 'min_samples_split': 5, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 18 with value: 6.43.
[I 2023-08-24 10:17:40,469] Trial 33 finished with value: 6.45 and parameters: {'n_estimators': 100, 'max_depth': 12, 'min_samples_split': 6, 'min_samples_leaf': 1, 'max_features': None, 'bootstrap': True}. Best is trial 18 with value: 6.43.
[I 2023-08-24 10:17:41,941] Trial 34 finished with value: 6.45 and parameters: {'n_estimators': 138, 'max_depth': 15, 'min_samples_split': 4, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 18 with value: 6.43.
[I 2023-08-24 10:17:42,622] Trial 35 finished with value: 6.45 and parameters: {'n_estimators': 66, 'max_depth': 11, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 18 with value: 6.43.
[I 2023-08-24 10:17:43,719] Trial 36 finished with value: 6.47 and parameters: {'n_estimators': 104, 'max_depth': 13, 'min_samples_split': 5, 'min_samples_leaf': 1, 'max_features': None, 'bootstrap': True}. Best is trial 18 with value: 6.43.
[I 2023-08-24 10:17:44,099] Trial 37 finished with value: 6.51 and parameters: {'n_estimators': 33, 'max_depth': 16, 'min_samples_split': 4, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 18 with value: 6.43.
[I 2023-08-24 10:17:45,944] Trial 38 finished with value: 6.44 and parameters: {'n_estimators': 222, 'max_depth': 10, 'min_samples_split': 6, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 18 with value: 6.43.
[I 2023-08-24 10:17:47,793] Trial 39 finished with value: 6.45 and parameters: {'n_estimators': 181, 'max_depth': 14, 'min_samples_split': 5, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 18 with value: 6.43.
[I 2023-08-24 10:17:48,662] Trial 40 finished with value: 6.45 and parameters: {'n_estimators': 82, 'max_depth': 23, 'min_samples_split': 4, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 18 with value: 6.43.
[I 2023-08-24 10:17:49,359] Trial 41 finished with value: 6.44 and parameters: {'n_estimators': 67, 'max_depth': 12, 'min_samples_split': 5, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 18 with value: 6.43.
[I 2023-08-24 10:17:49,825] Trial 42 finished with value: 6.49 and parameters: {'n_estimators': 40, 'max_depth': 12, 'min_samples_split': 5, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 18 with value: 6.43.
[I 2023-08-24 10:17:51,361] Trial 43 finished with value: 6.45 and parameters: {'n_estimators': 160, 'max_depth': 13, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 18 with value: 6.43.
[I 2023-08-24 10:17:52,425] Trial 44 finished with value: 6.45 and parameters: {'n_estimators': 110, 'max_depth': 11, 'min_samples_split': 5, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 18 with value: 6.43.
[I 2023-08-24 10:17:52,740] Trial 45 finished with value: 6.55 and parameters: {'n_estimators': 27, 'max_depth': 19, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 18 with value: 6.43.
[I 2023-08-24 10:17:53,440] Trial 46 finished with value: 6.45 and parameters: {'n_estimators': 71, 'max_depth': 13, 'min_samples_split': 5, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 18 with value: 6.43.
[I 2023-08-24 10:17:54,925] Trial 47 finished with value: 6.47 and parameters: {'n_estimators': 143, 'max_depth': 14, 'min_samples_split': 4, 'min_samples_leaf': 1, 'max_features': None, 'bootstrap': True}. Best is trial 18 with value: 6.43.
[I 2023-08-24 10:17:55,824] Trial 48 finished with value: 6.43 and parameters: {'n_estimators': 92, 'max_depth': 17, 'min_samples_split': 3, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 18 with value: 6.43.
[I 2023-08-24 10:17:56,391] Trial 49 finished with value: 6.44 and parameters: {'n_estimators': 60, 'max_depth': 10, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 18 with value: 6.43.

Лучшие параметры: {'n_estimators': 85, 'max_depth': 12, 'min_samples_split': 4, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}
Лучшее значение метрики: 6.43

<timed exec>:56: ExperimentalWarning:

plot_param_importances is experimental (supported from v2.2.0). The interface can change in the future.

<timed exec>:72: ExperimentalWarning:

plot_optimization_history is experimental (supported from v2.2.0). The interface can change in the future.

Wall time: 1min 13s


# Создаём график контура
fig = optuna.visualization.plot_contour(
    study, params=["max_depth", "n_estimators"])
fig.show()


%%time

def objective(trial):
    # Определяем параметры для подбора
    n_estimators = trial.suggest_int('n_estimators', 150, 355)
    max_depth = trial.suggest_int('max_depth', 12, 26)
    min_samples_split = trial.suggest_int('min_samples_split', 3, 6)
    min_samples_leaf = trial.suggest_int('min_samples_leaf', 2, 3)
    max_features = trial.suggest_categorical('max_features', [None])
    bootstrap = trial.suggest_categorical('bootstrap', [True])

# Создадём модель случайного леса с определенными параметрами
    model = RandomForestRegressor(n_estimators=n_estimators,
                                  max_depth=max_depth,
                                  min_samples_split=min_samples_split,
                                  min_samples_leaf=min_samples_leaf,
                                  max_features=max_features,
                                  bootstrap=bootstrap,
                                  random_state=RANDOM_STATE,  
                                  n_jobs=-1)

# Оцениваем качество модели с помощью кросс-валидации
    score = round(abs(cross_val_score(model,
                                      features_train,
                                      target_train,
                                      cv=kf,
                                      scoring='neg_mean_absolute_error', n_jobs=-1)).mean(), 2)
# Возврщаем значения метрики
    return score.mean()


# Создаём объект Study и запускаем оптимизацию
study = optuna.create_study(direction='minimize')
study.optimize(objective, n_trials=200)

print('Лучшие параметры:', study.best_params)
print('Лучшее значение метрики:', study.best_value)

# Создаём датафрейм с значениями гиперпараметров и метрикой
results = pd.DataFrame({'n_estimators': [],
                       'max_depth': [],
                        'min_samples_split': [],
                        'min_samples_leaf': [],
                        'max_features': [],
                        'bootstrap': [],
                        'score': []})
# Заполняем данными из best_trials
for trial in study.best_trials:
    params = trial.params
    params['score'] = trial.value
    results = results.append(params, ignore_index=True)
# Сортируем датафрейм по убыванию значения метрики
results = results.sort_values(by='score', ascending=False)
# Выводим 10 лучших моделей
display(results.head(10))

# Создаём график важности параметров
optuna.visualization.matplotlib.plot_param_importances(study)
plt.xlabel('Важность гиперпараметров',
           fontsize=12,
           color='DarkSlateGray')
plt.ylabel('Гиперпараметры',
           fontsize=12,
           color='DarkSlateGray')
plt.suptitle('Важность гиперпараметров',
             fontsize=15,
             color='DarkSlateGray')
plt.minorticks_on()
plt.grid(which='minor',
         linestyle=':')
plt.grid(True)

#  Создаём график истории оптимизации
optuna.visualization.matplotlib.plot_optimization_history(study)
plt.tight_layout()
plt.xlabel('Итерации',
           fontsize=12,
           color='DarkSlateGray')
plt.ylabel('Целевая метрика',
           fontsize=12,
           color='DarkSlateGray')
plt.suptitle('История оптимизации \n \n',
             fontsize=15,
             color='DarkSlateGray')
plt.minorticks_on()
plt.grid(which='minor',
         linestyle=':')
plt.grid(True)
plt.show()

# Сохранение модели с наилучшими параметрами
forest = RandomForestRegressor(n_estimators=study.best_params['n_estimators'],
                               max_depth=study.best_params['max_depth'],
                               min_samples_split=study.best_params['min_samples_split'],
                               min_samples_leaf=study.best_params['min_samples_leaf'],
                               max_features=study.best_params['max_features'],
                               bootstrap=study.best_params['bootstrap'],
                               random_state=RANDOM_STATE, n_jobs=-1)

[I 2023-08-24 10:17:58,391] A new study created in memory with name: no-name-9db2dd26-faab-49fa-8f6d-24d21c2b734b
[I 2023-08-24 10:18:01,598] Trial 0 finished with value: 6.44 and parameters: {'n_estimators': 319, 'max_depth': 25, 'min_samples_split': 5, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:18:03,801] Trial 1 finished with value: 6.44 and parameters: {'n_estimators': 233, 'max_depth': 21, 'min_samples_split': 6, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:18:05,271] Trial 2 finished with value: 6.44 and parameters: {'n_estimators': 158, 'max_depth': 20, 'min_samples_split': 4, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:18:07,962] Trial 3 finished with value: 6.44 and parameters: {'n_estimators': 281, 'max_depth': 17, 'min_samples_split': 3, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:18:10,470] Trial 4 finished with value: 6.44 and parameters: {'n_estimators': 249, 'max_depth': 26, 'min_samples_split': 5, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:18:13,373] Trial 5 finished with value: 6.44 and parameters: {'n_estimators': 315, 'max_depth': 26, 'min_samples_split': 4, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:18:16,190] Trial 6 finished with value: 6.44 and parameters: {'n_estimators': 312, 'max_depth': 12, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:18:18,435] Trial 7 finished with value: 6.45 and parameters: {'n_estimators': 199, 'max_depth': 20, 'min_samples_split': 5, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:18:21,021] Trial 8 finished with value: 6.44 and parameters: {'n_estimators': 243, 'max_depth': 22, 'min_samples_split': 3, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:18:23,629] Trial 9 finished with value: 6.44 and parameters: {'n_estimators': 279, 'max_depth': 13, 'min_samples_split': 5, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:18:26,932] Trial 10 finished with value: 6.45 and parameters: {'n_estimators': 355, 'max_depth': 24, 'min_samples_split': 6, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:18:28,911] Trial 11 finished with value: 6.44 and parameters: {'n_estimators': 205, 'max_depth': 17, 'min_samples_split': 6, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:18:31,063] Trial 12 finished with value: 6.45 and parameters: {'n_estimators': 211, 'max_depth': 23, 'min_samples_split': 6, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:18:34,484] Trial 13 finished with value: 6.44 and parameters: {'n_estimators': 355, 'max_depth': 22, 'min_samples_split': 5, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:18:37,155] Trial 14 finished with value: 6.44 and parameters: {'n_estimators': 284, 'max_depth': 17, 'min_samples_split': 4, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:18:39,344] Trial 15 finished with value: 6.45 and parameters: {'n_estimators': 222, 'max_depth': 24, 'min_samples_split': 6, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:18:41,374] Trial 16 finished with value: 6.44 and parameters: {'n_estimators': 177, 'max_depth': 19, 'min_samples_split': 5, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:18:44,741] Trial 17 finished with value: 6.44 and parameters: {'n_estimators': 312, 'max_depth': 25, 'min_samples_split': 6, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:18:47,420] Trial 18 finished with value: 6.44 and parameters: {'n_estimators': 265, 'max_depth': 21, 'min_samples_split': 5, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:18:49,782] Trial 19 finished with value: 6.44 and parameters: {'n_estimators': 233, 'max_depth': 15, 'min_samples_split': 4, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:18:53,228] Trial 20 finished with value: 6.45 and parameters: {'n_estimators': 337, 'max_depth': 19, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:18:54,782] Trial 21 finished with value: 6.44 and parameters: {'n_estimators': 158, 'max_depth': 20, 'min_samples_split': 4, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:18:56,328] Trial 22 finished with value: 6.45 and parameters: {'n_estimators': 150, 'max_depth': 22, 'min_samples_split': 4, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:18:58,091] Trial 23 finished with value: 6.44 and parameters: {'n_estimators': 185, 'max_depth': 18, 'min_samples_split': 3, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:18:59,830] Trial 24 finished with value: 6.44 and parameters: {'n_estimators': 176, 'max_depth': 24, 'min_samples_split': 4, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:19:02,327] Trial 25 finished with value: 6.44 and parameters: {'n_estimators': 259, 'max_depth': 15, 'min_samples_split': 5, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:19:04,556] Trial 26 finished with value: 6.44 and parameters: {'n_estimators': 225, 'max_depth': 20, 'min_samples_split': 4, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:19:07,367] Trial 27 finished with value: 6.44 and parameters: {'n_estimators': 299, 'max_depth': 21, 'min_samples_split': 5, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:19:10,516] Trial 28 finished with value: 6.45 and parameters: {'n_estimators': 334, 'max_depth': 23, 'min_samples_split': 3, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:19:13,263] Trial 29 finished with value: 6.44 and parameters: {'n_estimators': 282, 'max_depth': 18, 'min_samples_split': 3, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:19:15,170] Trial 30 finished with value: 6.44 and parameters: {'n_estimators': 195, 'max_depth': 15, 'min_samples_split': 5, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:19:17,903] Trial 31 finished with value: 6.44 and parameters: {'n_estimators': 267, 'max_depth': 16, 'min_samples_split': 3, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:19:21,041] Trial 32 finished with value: 6.44 and parameters: {'n_estimators': 303, 'max_depth': 26, 'min_samples_split': 3, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:19:23,572] Trial 33 finished with value: 6.44 and parameters: {'n_estimators': 246, 'max_depth': 18, 'min_samples_split': 4, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:19:26,956] Trial 34 finished with value: 6.45 and parameters: {'n_estimators': 325, 'max_depth': 19, 'min_samples_split': 4, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:19:29,828] Trial 35 finished with value: 6.44 and parameters: {'n_estimators': 303, 'max_depth': 26, 'min_samples_split': 3, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:19:32,834] Trial 36 finished with value: 6.44 and parameters: {'n_estimators': 295, 'max_depth': 21, 'min_samples_split': 5, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:19:35,736] Trial 37 finished with value: 6.44 and parameters: {'n_estimators': 325, 'max_depth': 13, 'min_samples_split': 6, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:19:38,554] Trial 38 finished with value: 6.44 and parameters: {'n_estimators': 273, 'max_depth': 17, 'min_samples_split': 4, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:19:40,971] Trial 39 finished with value: 6.44 and parameters: {'n_estimators': 252, 'max_depth': 23, 'min_samples_split': 6, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:19:43,248] Trial 40 finished with value: 6.45 and parameters: {'n_estimators': 239, 'max_depth': 25, 'min_samples_split': 3, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:19:45,479] Trial 41 finished with value: 6.44 and parameters: {'n_estimators': 216, 'max_depth': 25, 'min_samples_split': 5, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:19:48,374] Trial 42 finished with value: 6.45 and parameters: {'n_estimators': 286, 'max_depth': 26, 'min_samples_split': 5, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:19:51,001] Trial 43 finished with value: 6.44 and parameters: {'n_estimators': 255, 'max_depth': 24, 'min_samples_split': 5, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:19:53,340] Trial 44 finished with value: 6.44 and parameters: {'n_estimators': 232, 'max_depth': 22, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:19:56,534] Trial 45 finished with value: 6.44 and parameters: {'n_estimators': 345, 'max_depth': 16, 'min_samples_split': 5, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:19:58,612] Trial 46 finished with value: 6.45 and parameters: {'n_estimators': 201, 'max_depth': 25, 'min_samples_split': 5, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:20:01,297] Trial 47 finished with value: 6.44 and parameters: {'n_estimators': 275, 'max_depth': 20, 'min_samples_split': 6, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:20:04,148] Trial 48 finished with value: 6.44 and parameters: {'n_estimators': 292, 'max_depth': 22, 'min_samples_split': 5, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:20:07,163] Trial 49 finished with value: 6.44 and parameters: {'n_estimators': 314, 'max_depth': 23, 'min_samples_split': 4, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:20:09,450] Trial 50 finished with value: 6.45 and parameters: {'n_estimators': 245, 'max_depth': 19, 'min_samples_split': 6, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:20:12,449] Trial 51 finished with value: 6.44 and parameters: {'n_estimators': 318, 'max_depth': 26, 'min_samples_split': 4, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:20:15,654] Trial 52 finished with value: 6.45 and parameters: {'n_estimators': 344, 'max_depth': 25, 'min_samples_split': 4, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:20:18,148] Trial 53 finished with value: 6.44 and parameters: {'n_estimators': 262, 'max_depth': 24, 'min_samples_split': 4, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:20:21,129] Trial 54 finished with value: 6.44 and parameters: {'n_estimators': 307, 'max_depth': 26, 'min_samples_split': 4, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:20:22,732] Trial 55 finished with value: 6.44 and parameters: {'n_estimators': 161, 'max_depth': 21, 'min_samples_split': 5, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:20:24,726] Trial 56 finished with value: 6.45 and parameters: {'n_estimators': 210, 'max_depth': 24, 'min_samples_split': 4, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:20:27,848] Trial 57 finished with value: 6.44 and parameters: {'n_estimators': 323, 'max_depth': 16, 'min_samples_split': 6, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:20:30,557] Trial 58 finished with value: 6.44 and parameters: {'n_estimators': 290, 'max_depth': 25, 'min_samples_split': 5, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:20:32,171] Trial 59 finished with value: 6.44 and parameters: {'n_estimators': 173, 'max_depth': 13, 'min_samples_split': 3, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:20:35,282] Trial 60 finished with value: 6.44 and parameters: {'n_estimators': 332, 'max_depth': 18, 'min_samples_split': 5, 'min_samples_leaf': 3, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:20:38,121] Trial 61 finished with value: 6.44 and parameters: {'n_estimators': 311, 'max_depth': 12, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 0 with value: 6.44.
[I 2023-08-24 10:20:40,696] Trial 62 finished with value: 6.43 and parameters: {'n_estimators': 272, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:20:43,504] Trial 63 finished with value: 6.43 and parameters: {'n_estimators': 268, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:20:46,204] Trial 64 finished with value: 6.43 and parameters: {'n_estimators': 273, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:20:48,715] Trial 65 finished with value: 6.44 and parameters: {'n_estimators': 270, 'max_depth': 13, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:20:51,406] Trial 66 finished with value: 6.43 and parameters: {'n_estimators': 278, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:20:54,169] Trial 67 finished with value: 6.43 and parameters: {'n_estimators': 277, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:20:56,912] Trial 68 finished with value: 6.43 and parameters: {'n_estimators': 275, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:20:59,614] Trial 69 finished with value: 6.43 and parameters: {'n_estimators': 277, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:21:02,273] Trial 70 finished with value: 6.43 and parameters: {'n_estimators': 277, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:21:05,048] Trial 71 finished with value: 6.43 and parameters: {'n_estimators': 278, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:21:07,584] Trial 72 finished with value: 6.43 and parameters: {'n_estimators': 258, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:21:10,131] Trial 73 finished with value: 6.44 and parameters: {'n_estimators': 267, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:21:12,934] Trial 74 finished with value: 6.44 and parameters: {'n_estimators': 284, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:21:15,765] Trial 75 finished with value: 6.44 and parameters: {'n_estimators': 276, 'max_depth': 15, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:21:18,428] Trial 76 finished with value: 6.44 and parameters: {'n_estimators': 289, 'max_depth': 12, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:21:20,963] Trial 77 finished with value: 6.44 and parameters: {'n_estimators': 263, 'max_depth': 13, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:21:24,024] Trial 78 finished with value: 6.44 and parameters: {'n_estimators': 296, 'max_depth': 15, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:21:26,865] Trial 79 finished with value: 6.44 and parameters: {'n_estimators': 281, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:21:29,660] Trial 80 finished with value: 6.44 and parameters: {'n_estimators': 271, 'max_depth': 15, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:21:32,396] Trial 81 finished with value: 6.43 and parameters: {'n_estimators': 277, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:21:35,096] Trial 82 finished with value: 6.43 and parameters: {'n_estimators': 280, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:21:37,498] Trial 83 finished with value: 6.44 and parameters: {'n_estimators': 250, 'max_depth': 13, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:21:39,956] Trial 84 finished with value: 6.44 and parameters: {'n_estimators': 269, 'max_depth': 12, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:21:42,405] Trial 85 finished with value: 6.43 and parameters: {'n_estimators': 256, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:21:45,268] Trial 86 finished with value: 6.44 and parameters: {'n_estimators': 287, 'max_depth': 16, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:21:47,824] Trial 87 finished with value: 6.44 and parameters: {'n_estimators': 263, 'max_depth': 15, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:21:50,586] Trial 88 finished with value: 6.44 and parameters: {'n_estimators': 300, 'max_depth': 13, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:21:53,123] Trial 89 finished with value: 6.44 and parameters: {'n_estimators': 273, 'max_depth': 15, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:21:55,809] Trial 90 finished with value: 6.43 and parameters: {'n_estimators': 279, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:21:58,263] Trial 91 finished with value: 6.43 and parameters: {'n_estimators': 256, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:22:00,742] Trial 92 finished with value: 6.43 and parameters: {'n_estimators': 260, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:22:03,020] Trial 93 finished with value: 6.44 and parameters: {'n_estimators': 241, 'max_depth': 13, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:22:05,832] Trial 94 finished with value: 6.43 and parameters: {'n_estimators': 293, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:22:08,552] Trial 95 finished with value: 6.44 and parameters: {'n_estimators': 266, 'max_depth': 16, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:22:11,296] Trial 96 finished with value: 6.44 and parameters: {'n_estimators': 284, 'max_depth': 15, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:22:13,892] Trial 97 finished with value: 6.44 and parameters: {'n_estimators': 276, 'max_depth': 13, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:22:16,365] Trial 98 finished with value: 6.44 and parameters: {'n_estimators': 259, 'max_depth': 13, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:22:19,016] Trial 99 finished with value: 6.44 and parameters: {'n_estimators': 273, 'max_depth': 15, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:22:21,369] Trial 100 finished with value: 6.44 and parameters: {'n_estimators': 252, 'max_depth': 12, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:22:23,995] Trial 101 finished with value: 6.43 and parameters: {'n_estimators': 276, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:22:26,664] Trial 102 finished with value: 6.43 and parameters: {'n_estimators': 279, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:22:29,229] Trial 103 finished with value: 6.43 and parameters: {'n_estimators': 268, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:22:31,912] Trial 104 finished with value: 6.44 and parameters: {'n_estimators': 289, 'max_depth': 15, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:22:34,585] Trial 105 finished with value: 6.44 and parameters: {'n_estimators': 283, 'max_depth': 13, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:22:37,133] Trial 106 finished with value: 6.44 and parameters: {'n_estimators': 265, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:22:39,779] Trial 107 finished with value: 6.43 and parameters: {'n_estimators': 278, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:22:42,440] Trial 108 finished with value: 6.44 and parameters: {'n_estimators': 273, 'max_depth': 15, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:22:45,208] Trial 109 finished with value: 6.44 and parameters: {'n_estimators': 297, 'max_depth': 13, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:22:47,546] Trial 110 finished with value: 6.44 and parameters: {'n_estimators': 237, 'max_depth': 15, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:22:50,246] Trial 111 finished with value: 6.44 and parameters: {'n_estimators': 281, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:22:52,904] Trial 112 finished with value: 6.43 and parameters: {'n_estimators': 269, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:22:55,994] Trial 113 finished with value: 6.44 and parameters: {'n_estimators': 285, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:22:58,551] Trial 114 finished with value: 6.44 and parameters: {'n_estimators': 260, 'max_depth': 13, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:23:01,532] Trial 115 finished with value: 6.44 and parameters: {'n_estimators': 293, 'max_depth': 15, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:23:03,924] Trial 116 finished with value: 6.43 and parameters: {'n_estimators': 248, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:23:06,669] Trial 117 finished with value: 6.44 and parameters: {'n_estimators': 279, 'max_depth': 16, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:23:09,351] Trial 118 finished with value: 6.44 and parameters: {'n_estimators': 287, 'max_depth': 12, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:23:11,901] Trial 119 finished with value: 6.44 and parameters: {'n_estimators': 270, 'max_depth': 13, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:23:14,341] Trial 120 finished with value: 6.43 and parameters: {'n_estimators': 255, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:23:16,800] Trial 121 finished with value: 6.43 and parameters: {'n_estimators': 256, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:23:19,354] Trial 122 finished with value: 6.44 and parameters: {'n_estimators': 266, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:23:22,152] Trial 123 finished with value: 6.44 and parameters: {'n_estimators': 275, 'max_depth': 15, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:23:24,607] Trial 124 finished with value: 6.44 and parameters: {'n_estimators': 263, 'max_depth': 13, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:23:27,310] Trial 125 finished with value: 6.43 and parameters: {'n_estimators': 273, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:23:30,105] Trial 126 finished with value: 6.44 and parameters: {'n_estimators': 283, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:23:32,813] Trial 127 finished with value: 6.44 and parameters: {'n_estimators': 279, 'max_depth': 15, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:23:35,244] Trial 128 finished with value: 6.44 and parameters: {'n_estimators': 257, 'max_depth': 13, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:23:37,873] Trial 129 finished with value: 6.43 and parameters: {'n_estimators': 270, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:23:40,443] Trial 130 finished with value: 6.44 and parameters: {'n_estimators': 262, 'max_depth': 16, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:23:43,111] Trial 131 finished with value: 6.43 and parameters: {'n_estimators': 278, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:23:45,896] Trial 132 finished with value: 6.44 and parameters: {'n_estimators': 281, 'max_depth': 15, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:23:48,757] Trial 133 finished with value: 6.44 and parameters: {'n_estimators': 290, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:23:51,558] Trial 134 finished with value: 6.44 and parameters: {'n_estimators': 274, 'max_depth': 13, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:23:54,265] Trial 135 finished with value: 6.44 and parameters: {'n_estimators': 266, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:23:57,170] Trial 136 finished with value: 6.44 and parameters: {'n_estimators': 285, 'max_depth': 15, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:23:59,846] Trial 137 finished with value: 6.43 and parameters: {'n_estimators': 252, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:24:02,687] Trial 138 finished with value: 6.44 and parameters: {'n_estimators': 303, 'max_depth': 13, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:24:05,363] Trial 139 finished with value: 6.44 and parameters: {'n_estimators': 271, 'max_depth': 15, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:24:08,100] Trial 140 finished with value: 6.43 and parameters: {'n_estimators': 277, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:24:10,503] Trial 141 finished with value: 6.43 and parameters: {'n_estimators': 245, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:24:13,053] Trial 142 finished with value: 6.43 and parameters: {'n_estimators': 257, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:24:15,612] Trial 143 finished with value: 6.44 and parameters: {'n_estimators': 264, 'max_depth': 13, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:24:18,342] Trial 144 finished with value: 6.43 and parameters: {'n_estimators': 282, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:24:20,954] Trial 145 finished with value: 6.44 and parameters: {'n_estimators': 269, 'max_depth': 15, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:24:23,559] Trial 146 finished with value: 6.44 and parameters: {'n_estimators': 276, 'max_depth': 13, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:24:26,029] Trial 147 finished with value: 6.43 and parameters: {'n_estimators': 259, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:24:28,748] Trial 148 finished with value: 6.43 and parameters: {'n_estimators': 287, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:24:31,233] Trial 149 finished with value: 6.44 and parameters: {'n_estimators': 249, 'max_depth': 15, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:24:33,794] Trial 150 finished with value: 6.44 and parameters: {'n_estimators': 273, 'max_depth': 13, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:24:36,356] Trial 151 finished with value: 6.43 and parameters: {'n_estimators': 260, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:24:38,960] Trial 152 finished with value: 6.44 and parameters: {'n_estimators': 266, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:24:41,472] Trial 153 finished with value: 6.43 and parameters: {'n_estimators': 254, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:24:44,220] Trial 154 finished with value: 6.43 and parameters: {'n_estimators': 280, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:24:46,850] Trial 155 finished with value: 6.44 and parameters: {'n_estimators': 260, 'max_depth': 15, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:24:49,504] Trial 156 finished with value: 6.44 and parameters: {'n_estimators': 269, 'max_depth': 13, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:24:52,191] Trial 157 finished with value: 6.43 and parameters: {'n_estimators': 275, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:24:54,998] Trial 158 finished with value: 6.44 and parameters: {'n_estimators': 290, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:24:57,586] Trial 159 finished with value: 6.44 and parameters: {'n_estimators': 264, 'max_depth': 15, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:25:00,288] Trial 160 finished with value: 6.43 and parameters: {'n_estimators': 280, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:25:03,061] Trial 161 finished with value: 6.43 and parameters: {'n_estimators': 286, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:25:05,902] Trial 162 finished with value: 6.44 and parameters: {'n_estimators': 296, 'max_depth': 13, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:25:08,700] Trial 163 finished with value: 6.43 and parameters: {'n_estimators': 277, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:25:11,319] Trial 164 finished with value: 6.43 and parameters: {'n_estimators': 272, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:25:14,151] Trial 165 finished with value: 6.44 and parameters: {'n_estimators': 284, 'max_depth': 15, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:25:16,990] Trial 166 finished with value: 6.43 and parameters: {'n_estimators': 291, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:25:19,519] Trial 167 finished with value: 6.44 and parameters: {'n_estimators': 268, 'max_depth': 13, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:25:21,978] Trial 168 finished with value: 6.44 and parameters: {'n_estimators': 253, 'max_depth': 15, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:25:24,495] Trial 169 finished with value: 6.44 and parameters: {'n_estimators': 261, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:25:27,002] Trial 170 finished with value: 6.44 and parameters: {'n_estimators': 271, 'max_depth': 13, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:25:29,778] Trial 171 finished with value: 6.43 and parameters: {'n_estimators': 277, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:25:32,653] Trial 172 finished with value: 6.43 and parameters: {'n_estimators': 293, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:25:35,409] Trial 173 finished with value: 6.43 and parameters: {'n_estimators': 275, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:25:38,239] Trial 174 finished with value: 6.44 and parameters: {'n_estimators': 281, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:25:40,820] Trial 175 finished with value: 6.43 and parameters: {'n_estimators': 268, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:25:43,371] Trial 176 finished with value: 6.44 and parameters: {'n_estimators': 264, 'max_depth': 15, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:25:46,136] Trial 177 finished with value: 6.44 and parameters: {'n_estimators': 283, 'max_depth': 13, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:25:48,732] Trial 178 finished with value: 6.43 and parameters: {'n_estimators': 274, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:25:51,492] Trial 179 finished with value: 6.44 and parameters: {'n_estimators': 277, 'max_depth': 15, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:25:54,662] Trial 180 finished with value: 6.44 and parameters: {'n_estimators': 307, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:25:57,507] Trial 181 finished with value: 6.43 and parameters: {'n_estimators': 279, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:26:00,159] Trial 182 finished with value: 6.43 and parameters: {'n_estimators': 272, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:26:02,951] Trial 183 finished with value: 6.43 and parameters: {'n_estimators': 287, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:26:05,648] Trial 184 finished with value: 6.44 and parameters: {'n_estimators': 280, 'max_depth': 13, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:26:08,267] Trial 185 finished with value: 6.43 and parameters: {'n_estimators': 256, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:26:10,906] Trial 186 finished with value: 6.44 and parameters: {'n_estimators': 266, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:26:13,784] Trial 187 finished with value: 6.44 and parameters: {'n_estimators': 284, 'max_depth': 15, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:26:16,438] Trial 188 finished with value: 6.44 and parameters: {'n_estimators': 271, 'max_depth': 13, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:26:19,182] Trial 189 finished with value: 6.43 and parameters: {'n_estimators': 277, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:26:21,690] Trial 190 finished with value: 6.44 and parameters: {'n_estimators': 249, 'max_depth': 15, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:26:24,366] Trial 191 finished with value: 6.43 and parameters: {'n_estimators': 268, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:26:26,780] Trial 192 finished with value: 6.44 and parameters: {'n_estimators': 261, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:26:29,617] Trial 193 finished with value: 6.43 and parameters: {'n_estimators': 274, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:26:32,330] Trial 194 finished with value: 6.43 and parameters: {'n_estimators': 279, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:26:34,891] Trial 195 finished with value: 6.44 and parameters: {'n_estimators': 267, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:26:37,356] Trial 196 finished with value: 6.44 and parameters: {'n_estimators': 258, 'max_depth': 13, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:26:39,992] Trial 197 finished with value: 6.43 and parameters: {'n_estimators': 273, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:26:42,650] Trial 198 finished with value: 6.44 and parameters: {'n_estimators': 263, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.
[I 2023-08-24 10:26:45,488] Trial 199 finished with value: 6.44 and parameters: {'n_estimators': 283, 'max_depth': 15, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}. Best is trial 62 with value: 6.43.

Лучшие параметры: {'n_estimators': 272, 'max_depth': 14, 'min_samples_split': 6, 'min_samples_leaf': 2, 'max_features': None, 'bootstrap': True}
Лучшее значение метрики: 6.43

<timed exec>:56: ExperimentalWarning:

plot_param_importances is experimental (supported from v2.2.0). The interface can change in the future.

<timed exec>:72: ExperimentalWarning:

plot_optimization_history is experimental (supported from v2.2.0). The interface can change in the future.

Wall time: 8min 49s


# Создаём график контура
fig = optuna.visualization.plot_contour(study, params=["max_depth", "n_estimators"])
fig.show()


forest

RandomForestRegressor(max_depth=14, max_features=None, min_samples_leaf=2,
                      min_samples_split=6, n_estimators=272, n_jobs=-1,
                      random_state=140823)


%%time

def objective(trial):
    # Определяем параметры для подбора
    n_estimators = trial.suggest_int('n_estimators', 10, 500)
    max_depth = trial.suggest_int('max_depth', 2, 30)
    min_child_samples = trial.suggest_int('min_child_samples', 1, 8)
    learning_rate = trial.suggest_float('learning_rate', 0.01, 0.1, step=0.01)
    subsample = trial.suggest_float('subsample', 0.5, 1.0, step=0.01)
    colsample_bytree = trial.suggest_float('colsample_bytree', 0.5, 1.0, step=0.01)
    
    # Создаем модель с определенными параметрами
    model = LGBMRegressor(n_estimators=n_estimators,
                          max_depth=max_depth,
                          min_child_samples=min_child_samples,
                          learning_rate=learning_rate,
                          subsample=subsample,
                          colsample_bytree=colsample_bytree,
                          random_state=RANDOM_STATE)
    
    # Оцениваем качество модели с помощью кросс-валидации
    score = round(abs(cross_val_score(model,
                                      features_train,
                                      target_train,
                                      cv=kf,
                                      scoring='neg_mean_absolute_error', n_jobs=-1)).mean(), 2)
    
    # Возвращаем значения метрики
    return score.mean()

# Создаем объект Study и запускаем оптимизацию
study = optuna.create_study(direction='minimize')
study.optimize(objective, n_trials=50)

print('Лучшие параметры:', study.best_params)
print('Лучшее значение метрики:', study.best_value)

# Создаем датафрейм с значениями гиперпараметров и метрикой
results = pd.DataFrame({'n_estimators': [],
                       'max_depth': [],
                       'min_child_samples': [],
                       'learning_rate': [],
                       'subsample': [],
                       'colsample_bytree': [],
                       'score': []})
# Заполняем данными из best_trials
for trial in study.best_trials:
    params = trial.params
    params['score'] = trial.value
    results = results.append(params, ignore_index=True)

# Сортируем датафрейм по убыванию значения метрики
results = results.sort_values(by='score', ascending=False)

# Выводим 10 лучших моделей
display(results.head(10))

# Создаем график важности параметров
optuna.visualization.matplotlib.plot_param_importances(study)
plt.xlabel('Важность гиперпараметров',
           fontsize=12,
           color='DarkSlateGray')
plt.ylabel('Гиперпараметры',
           fontsize=12,
           color='DarkSlateGray')
plt.suptitle('Важность гиперпараметров',
             fontsize=15,
             color='DarkSlateGray')
plt.minorticks_on()
plt.grid(which='minor',
         linestyle=':')
plt.grid(True)

# Создаем график истории оптимизации
optuna.visualization.matplotlib.plot_optimization_history(study)
plt.tight_layout()
plt.xlabel('Итерации',
           fontsize=12,
           color='DarkSlateGray')
plt.ylabel('Целевая метрика',
           fontsize=12,
           color='DarkSlateGray')
plt.suptitle('История оптимизации \n \n',
             fontsize=15,
             color='DarkSlateGray')
plt.minorticks_on()
plt.grid(which='minor',
         linestyle=':')
plt.grid(True)
plt.show()

[I 2023-08-24 10:26:48,440] A new study created in memory with name: no-name-7dc85ea7-653b-4253-aba0-5241b419d609
[I 2023-08-24 10:26:49,709] Trial 0 finished with value: 6.49 and parameters: {'n_estimators': 494, 'max_depth': 20, 'min_child_samples': 7, 'learning_rate': 0.060000000000000005, 'subsample': 0.8200000000000001, 'colsample_bytree': 0.75}. Best is trial 0 with value: 6.49.
[I 2023-08-24 10:26:50,844] Trial 1 finished with value: 6.53 and parameters: {'n_estimators': 450, 'max_depth': 15, 'min_child_samples': 3, 'learning_rate': 0.04, 'subsample': 0.9299999999999999, 'colsample_bytree': 0.8200000000000001}. Best is trial 0 with value: 6.49.
[I 2023-08-24 10:26:51,021] Trial 2 finished with value: 6.76 and parameters: {'n_estimators': 58, 'max_depth': 9, 'min_child_samples': 7, 'learning_rate': 0.02, 'subsample': 0.5, 'colsample_bytree': 0.86}. Best is trial 0 with value: 6.49.
[I 2023-08-24 10:26:52,065] Trial 3 finished with value: 6.55 and parameters: {'n_estimators': 448, 'max_depth': 7, 'min_child_samples': 2, 'learning_rate': 0.060000000000000005, 'subsample': 0.6799999999999999, 'colsample_bytree': 0.9299999999999999}. Best is trial 0 with value: 6.49.
[I 2023-08-24 10:26:52,184] Trial 4 finished with value: 6.33 and parameters: {'n_estimators': 180, 'max_depth': 3, 'min_child_samples': 1, 'learning_rate': 0.05, 'subsample': 0.58, 'colsample_bytree': 0.51}. Best is trial 4 with value: 6.33.
[I 2023-08-24 10:26:52,642] Trial 5 finished with value: 6.57 and parameters: {'n_estimators': 195, 'max_depth': 11, 'min_child_samples': 5, 'learning_rate': 0.01, 'subsample': 0.5, 'colsample_bytree': 0.73}. Best is trial 4 with value: 6.33.
[I 2023-08-24 10:26:53,907] Trial 6 finished with value: 6.33 and parameters: {'n_estimators': 486, 'max_depth': 9, 'min_child_samples': 4, 'learning_rate': 0.01, 'subsample': 0.63, 'colsample_bytree': 0.81}. Best is trial 4 with value: 6.33.
[I 2023-08-24 10:26:54,167] Trial 7 finished with value: 6.4 and parameters: {'n_estimators': 89, 'max_depth': 17, 'min_child_samples': 7, 'learning_rate': 0.08, 'subsample': 0.8300000000000001, 'colsample_bytree': 0.91}. Best is trial 4 with value: 6.33.
[I 2023-08-24 10:26:54,931] Trial 8 finished with value: 6.36 and parameters: {'n_estimators': 301, 'max_depth': 25, 'min_child_samples': 7, 'learning_rate': 0.03, 'subsample': 0.9199999999999999, 'colsample_bytree': 0.9}. Best is trial 4 with value: 6.33.
[I 2023-08-24 10:26:55,032] Trial 9 finished with value: 7.1 and parameters: {'n_estimators': 167, 'max_depth': 2, 'min_child_samples': 6, 'learning_rate': 0.01, 'subsample': 0.87, 'colsample_bytree': 0.87}. Best is trial 4 with value: 6.33.
[I 2023-08-24 10:26:55,220] Trial 10 finished with value: 6.34 and parameters: {'n_estimators': 309, 'max_depth': 2, 'min_child_samples': 1, 'learning_rate': 0.09999999999999999, 'subsample': 0.63, 'colsample_bytree': 0.51}. Best is trial 4 with value: 6.33.
[I 2023-08-24 10:26:55,944] Trial 11 finished with value: 6.4 and parameters: {'n_estimators': 367, 'max_depth': 7, 'min_child_samples': 4, 'learning_rate': 0.04, 'subsample': 0.62, 'colsample_bytree': 0.5700000000000001}. Best is trial 4 with value: 6.33.
[I 2023-08-24 10:26:56,434] Trial 12 finished with value: 6.52 and parameters: {'n_estimators': 200, 'max_depth': 30, 'min_child_samples': 1, 'learning_rate': 0.08, 'subsample': 0.7, 'colsample_bytree': 0.64}. Best is trial 4 with value: 6.33.
[I 2023-08-24 10:26:56,884] Trial 13 finished with value: 6.4 and parameters: {'n_estimators': 138, 'max_depth': 12, 'min_child_samples': 3, 'learning_rate': 0.05, 'subsample': 0.58, 'colsample_bytree': 1.0}. Best is trial 4 with value: 6.33.
[I 2023-08-24 10:26:57,238] Trial 14 finished with value: 6.51 and parameters: {'n_estimators': 276, 'max_depth': 5, 'min_child_samples': 5, 'learning_rate': 0.08, 'subsample': 0.74, 'colsample_bytree': 0.65}. Best is trial 4 with value: 6.33.
[I 2023-08-24 10:26:57,331] Trial 15 finished with value: 7.33 and parameters: {'n_estimators': 14, 'max_depth': 5, 'min_child_samples': 3, 'learning_rate': 0.03, 'subsample': 0.55, 'colsample_bytree': 0.79}. Best is trial 4 with value: 6.33.
[I 2023-08-24 10:26:58,216] Trial 16 finished with value: 6.44 and parameters: {'n_estimators': 374, 'max_depth': 14, 'min_child_samples': 2, 'learning_rate': 0.06999999999999999, 'subsample': 0.64, 'colsample_bytree': 0.69}. Best is trial 4 with value: 6.33.
[I 2023-08-24 10:26:58,656] Trial 17 finished with value: 6.48 and parameters: {'n_estimators': 238, 'max_depth': 19, 'min_child_samples': 4, 'learning_rate': 0.09999999999999999, 'subsample': 0.5700000000000001, 'colsample_bytree': 0.58}. Best is trial 4 with value: 6.33.
[I 2023-08-24 10:26:59,369] Trial 18 finished with value: 6.37 and parameters: {'n_estimators': 380, 'max_depth': 10, 'min_child_samples': 8, 'learning_rate': 0.02, 'subsample': 1.0, 'colsample_bytree': 0.51}. Best is trial 4 with value: 6.33.
[I 2023-08-24 10:26:59,610] Trial 19 finished with value: 6.31 and parameters: {'n_estimators': 119, 'max_depth': 5, 'min_child_samples': 2, 'learning_rate': 0.05, 'subsample': 0.71, 'colsample_bytree': 0.8}. Best is trial 19 with value: 6.31.
[I 2023-08-24 10:26:59,775] Trial 20 finished with value: 6.32 and parameters: {'n_estimators': 107, 'max_depth': 4, 'min_child_samples': 2, 'learning_rate': 0.05, 'subsample': 0.77, 'colsample_bytree': 0.7}. Best is trial 19 with value: 6.31.
[I 2023-08-24 10:26:59,989] Trial 21 finished with value: 6.36 and parameters: {'n_estimators': 120, 'max_depth': 5, 'min_child_samples': 2, 'learning_rate': 0.05, 'subsample': 0.77, 'colsample_bytree': 0.59}. Best is trial 19 with value: 6.31.
[I 2023-08-24 10:27:00,084] Trial 22 finished with value: 6.87 and parameters: {'n_estimators': 67, 'max_depth': 2, 'min_child_samples': 1, 'learning_rate': 0.04, 'subsample': 0.75, 'colsample_bytree': 0.7}. Best is trial 19 with value: 6.31.
[I 2023-08-24 10:27:00,320] Trial 23 finished with value: 6.34 and parameters: {'n_estimators': 124, 'max_depth': 5, 'min_child_samples': 2, 'learning_rate': 0.060000000000000005, 'subsample': 0.71, 'colsample_bytree': 0.78}. Best is trial 19 with value: 6.31.
[I 2023-08-24 10:27:00,439] Trial 24 finished with value: 6.86 and parameters: {'n_estimators': 21, 'max_depth': 7, 'min_child_samples': 1, 'learning_rate': 0.05, 'subsample': 0.8, 'colsample_bytree': 0.63}. Best is trial 19 with value: 6.31.
[I 2023-08-24 10:27:00,674] Trial 25 finished with value: 6.39 and parameters: {'n_estimators': 229, 'max_depth': 4, 'min_child_samples': 3, 'learning_rate': 0.06999999999999999, 'subsample': 0.67, 'colsample_bytree': 0.7}. Best is trial 19 with value: 6.31.
[I 2023-08-24 10:27:01,149] Trial 26 finished with value: 6.43 and parameters: {'n_estimators': 161, 'max_depth': 13, 'min_child_samples': 2, 'learning_rate': 0.06999999999999999, 'subsample': 0.78, 'colsample_bytree': 0.97}. Best is trial 19 with value: 6.31.
[I 2023-08-24 10:27:01,400] Trial 27 finished with value: 6.48 and parameters: {'n_estimators': 92, 'max_depth': 8, 'min_child_samples': 1, 'learning_rate': 0.03, 'subsample': 0.8500000000000001, 'colsample_bytree': 0.54}. Best is trial 19 with value: 6.31.
[I 2023-08-24 10:27:01,569] Trial 28 finished with value: 6.38 and parameters: {'n_estimators': 170, 'max_depth': 3, 'min_child_samples': 2, 'learning_rate': 0.04, 'subsample': 0.73, 'colsample_bytree': 0.8400000000000001}. Best is trial 19 with value: 6.31.
[I 2023-08-24 10:27:01,745] Trial 29 finished with value: 6.53 and parameters: {'n_estimators': 42, 'max_depth': 22, 'min_child_samples': 3, 'learning_rate': 0.060000000000000005, 'subsample': 0.89, 'colsample_bytree': 0.66}. Best is trial 19 with value: 6.31.
[I 2023-08-24 10:27:01,923] Trial 30 finished with value: 6.35 and parameters: {'n_estimators': 95, 'max_depth': 4, 'min_child_samples': 1, 'learning_rate': 0.05, 'subsample': 0.54, 'colsample_bytree': 0.74}. Best is trial 19 with value: 6.31.
[I 2023-08-24 10:27:02,965] Trial 31 finished with value: 6.54 and parameters: {'n_estimators': 484, 'max_depth': 7, 'min_child_samples': 4, 'learning_rate': 0.09, 'subsample': 0.61, 'colsample_bytree': 0.8}. Best is trial 19 with value: 6.31.
[I 2023-08-24 10:27:03,478] Trial 32 finished with value: 6.4 and parameters: {'n_estimators': 192, 'max_depth': 9, 'min_child_samples': 5, 'learning_rate': 0.02, 'subsample': 0.66, 'colsample_bytree': 0.76}. Best is trial 19 with value: 6.31.
[I 2023-08-24 10:27:04,059] Trial 33 finished with value: 6.52 and parameters: {'n_estimators': 218, 'max_depth': 10, 'min_child_samples': 6, 'learning_rate': 0.060000000000000005, 'subsample': 0.6, 'colsample_bytree': 0.8500000000000001}. Best is trial 19 with value: 6.31.
[I 2023-08-24 10:27:04,438] Trial 34 finished with value: 6.41 and parameters: {'n_estimators': 127, 'max_depth': 7, 'min_child_samples': 3, 'learning_rate': 0.03, 'subsample': 0.6799999999999999, 'colsample_bytree': 0.8200000000000001}. Best is trial 19 with value: 6.31.
[I 2023-08-24 10:27:05,200] Trial 35 finished with value: 6.39 and parameters: {'n_estimators': 264, 'max_depth': 16, 'min_child_samples': 2, 'learning_rate': 0.04, 'subsample': 0.53, 'colsample_bytree': 0.88}. Best is trial 19 with value: 6.31.
[I 2023-08-24 10:27:06,126] Trial 36 finished with value: 6.3 and parameters: {'n_estimators': 417, 'max_depth': 6, 'min_child_samples': 4, 'learning_rate': 0.02, 'subsample': 0.71, 'colsample_bytree': 0.8200000000000001}. Best is trial 36 with value: 6.3.
[I 2023-08-24 10:27:06,534] Trial 37 finished with value: 6.41 and parameters: {'n_estimators': 414, 'max_depth': 4, 'min_child_samples': 2, 'learning_rate': 0.05, 'subsample': 0.8, 'colsample_bytree': 0.73}. Best is trial 36 with value: 6.3.
[I 2023-08-24 10:27:06,765] Trial 38 finished with value: 6.78 and parameters: {'n_estimators': 60, 'max_depth': 11, 'min_child_samples': 3, 'learning_rate': 0.02, 'subsample': 0.71, 'colsample_bytree': 0.77}. Best is trial 36 with value: 6.3.
[I 2023-08-24 10:27:07,404] Trial 39 finished with value: 6.57 and parameters: {'n_estimators': 332, 'max_depth': 6, 'min_child_samples': 6, 'learning_rate': 0.06999999999999999, 'subsample': 0.77, 'colsample_bytree': 0.9299999999999999}. Best is trial 36 with value: 6.3.
[I 2023-08-24 10:27:07,568] Trial 40 finished with value: 6.4 and parameters: {'n_estimators': 151, 'max_depth': 3, 'min_child_samples': 1, 'learning_rate': 0.04, 'subsample': 0.8300000000000001, 'colsample_bytree': 0.8300000000000001}. Best is trial 36 with value: 6.3.
[I 2023-08-24 10:27:08,839] Trial 41 finished with value: 6.33 and parameters: {'n_estimators': 494, 'max_depth': 9, 'min_child_samples': 4, 'learning_rate': 0.01, 'subsample': 0.64, 'colsample_bytree': 0.81}. Best is trial 36 with value: 6.3.
[I 2023-08-24 10:27:09,053] Trial 42 finished with value: 6.67 and parameters: {'n_estimators': 446, 'max_depth': 2, 'min_child_samples': 4, 'learning_rate': 0.01, 'subsample': 0.69, 'colsample_bytree': 0.61}. Best is trial 36 with value: 6.3.
[I 2023-08-24 10:27:09,931] Trial 43 finished with value: 6.36 and parameters: {'n_estimators': 426, 'max_depth': 6, 'min_child_samples': 5, 'learning_rate': 0.02, 'subsample': 0.59, 'colsample_bytree': 0.89}. Best is trial 36 with value: 6.3.
[I 2023-08-24 10:27:11,159] Trial 44 finished with value: 6.37 and parameters: {'n_estimators': 469, 'max_depth': 8, 'min_child_samples': 3, 'learning_rate': 0.01, 'subsample': 0.73, 'colsample_bytree': 0.86}. Best is trial 36 with value: 6.3.
[I 2023-08-24 10:27:11,299] Trial 45 finished with value: 6.82 and parameters: {'n_estimators': 95, 'max_depth': 3, 'min_child_samples': 4, 'learning_rate': 0.02, 'subsample': 0.66, 'colsample_bytree': 0.9199999999999999}. Best is trial 36 with value: 6.3.
[I 2023-08-24 10:27:11,796] Trial 46 finished with value: 6.38 and parameters: {'n_estimators': 188, 'max_depth': 29, 'min_child_samples': 5, 'learning_rate': 0.03, 'subsample': 0.51, 'colsample_bytree': 0.73}. Best is trial 36 with value: 6.3.
[I 2023-08-24 10:27:12,443] Trial 47 finished with value: 6.39 and parameters: {'n_estimators': 414, 'max_depth': 6, 'min_child_samples': 6, 'learning_rate': 0.03, 'subsample': 0.56, 'colsample_bytree': 0.55}. Best is trial 36 with value: 6.3.
[I 2023-08-24 10:27:13,499] Trial 48 finished with value: 6.47 and parameters: {'n_estimators': 458, 'max_depth': 8, 'min_child_samples': 2, 'learning_rate': 0.060000000000000005, 'subsample': 0.63, 'colsample_bytree': 0.76}. Best is trial 36 with value: 6.3.
[I 2023-08-24 10:27:14,383] Trial 49 finished with value: 6.52 and parameters: {'n_estimators': 333, 'max_depth': 12, 'min_child_samples': 3, 'learning_rate': 0.05, 'subsample': 0.72, 'colsample_bytree': 0.95}. Best is trial 36 with value: 6.3.

Лучшие параметры: {'n_estimators': 417, 'max_depth': 6, 'min_child_samples': 4, 'learning_rate': 0.02, 'subsample': 0.71, 'colsample_bytree': 0.8200000000000001}
Лучшее значение метрики: 6.3

<timed exec>:57: ExperimentalWarning:

plot_param_importances is experimental (supported from v2.2.0). The interface can change in the future.

<timed exec>:73: ExperimentalWarning:

plot_optimization_history is experimental (supported from v2.2.0). The interface can change in the future.

Wall time: 28.6 s


# Создаём график контура
fig = optuna.visualization.plot_contour(study, params=["max_depth", "n_estimators"])
fig.show()


%%time

def objective(trial):
    # Определяем параметры для подбора
    n_estimators = trial.suggest_int('n_estimators', 100, 400)
    max_depth = trial.suggest_int('max_depth', 1, 10)
    min_child_samples = trial.suggest_int('min_child_samples', 6, 10)
    learning_rate = trial.suggest_float('learning_rate', 0.01, 0.1, step=0.01)
    subsample = trial.suggest_float('subsample', 0.5, 0.9, step=0.01)
    colsample_bytree = trial.suggest_float('colsample_bytree', 0.7, 1, step=0.1)
    
    # Создаем модель с определенными параметрами
    model = LGBMRegressor(n_estimators=n_estimators,
                          max_depth=max_depth,
                          min_child_samples=min_child_samples,
                          learning_rate=learning_rate,
                          subsample=subsample,
                          colsample_bytree=colsample_bytree,
                          random_state=RANDOM_STATE)
    
    # Оцениваем качество модели с помощью кросс-валидации
    score = round(abs(cross_val_score(model,
                                      features_train,
                                      target_train,
                                      cv=kf,
                                      scoring='neg_mean_absolute_error', n_jobs=-1)).mean(), 2)
    
    # Возвращаем значения метрики
    return score.mean()

# Создаем объект Study и запускаем оптимизацию
study = optuna.create_study(direction='minimize')
study.optimize(objective, n_trials=50)

print('Лучшие параметры:', study.best_params)
print('Лучшее значение метрики:', study.best_value)

# Создаем датафрейм с значениями гиперпараметров и метрикой
results = pd.DataFrame({'n_estimators': [],
                       'max_depth': [],
                       'min_child_samples': [],
                       'learning_rate': [],
                       'subsample': [],
                       'colsample_bytree': [],
                       'score': []})
# Заполняем данными из best_trials
for trial in study.best_trials:
    params = trial.params
    params['score'] = trial.value
    results = results.append(params, ignore_index=True)

# Сортируем датафрейм по убыванию значения метрики
results = results.sort_values(by='score', ascending=False)

# Выводим 10 лучших моделей
display(results.head(10))

# Создаем график важности параметров
optuna.visualization.matplotlib.plot_param_importances(study)
plt.xlabel('Важность гиперпараметров',
           fontsize=12,
           color='DarkSlateGray')
plt.ylabel('Гиперпараметры',
           fontsize=12,
           color='DarkSlateGray')
plt.suptitle('Важность гиперпараметров',
             fontsize=15,
             color='DarkSlateGray')
plt.minorticks_on()
plt.grid(which='minor',
         linestyle=':')
plt.grid(True)

# Создаем график истории оптимизации
optuna.visualization.matplotlib.plot_optimization_history(study)
plt.tight_layout()
plt.xlabel('Итерации',
           fontsize=12,
           color='DarkSlateGray')
plt.ylabel('Целевая метрика',
           fontsize=12,
           color='DarkSlateGray')
plt.suptitle('История оптимизации \n \n',
             fontsize=15,
             color='DarkSlateGray')
plt.minorticks_on()
plt.grid(which='minor',
         linestyle=':')
plt.grid(True)
plt.show()

[I 2023-08-24 10:27:17,065] A new study created in memory with name: no-name-b9e8db1d-6e83-4904-b89d-56697bd66880
[I 2023-08-24 10:27:17,198] Trial 0 finished with value: 6.35 and parameters: {'n_estimators': 120, 'max_depth': 4, 'min_child_samples': 8, 'learning_rate': 0.04, 'subsample': 0.8400000000000001, 'colsample_bytree': 0.7999999999999999}. Best is trial 0 with value: 6.35.
[I 2023-08-24 10:27:17,531] Trial 1 finished with value: 6.41 and parameters: {'n_estimators': 241, 'max_depth': 5, 'min_child_samples': 7, 'learning_rate': 0.08, 'subsample': 0.59, 'colsample_bytree': 0.7}. Best is trial 0 with value: 6.35.
[I 2023-08-24 10:27:18,420] Trial 2 finished with value: 6.56 and parameters: {'n_estimators': 372, 'max_depth': 9, 'min_child_samples': 8, 'learning_rate': 0.08, 'subsample': 0.6799999999999999, 'colsample_bytree': 1.0}. Best is trial 0 with value: 6.35.
[I 2023-08-24 10:27:18,669] Trial 3 finished with value: 6.57 and parameters: {'n_estimators': 101, 'max_depth': 10, 'min_child_samples': 7, 'learning_rate': 0.02, 'subsample': 0.88, 'colsample_bytree': 0.7}. Best is trial 0 with value: 6.35.
[I 2023-08-24 10:27:18,753] Trial 4 finished with value: 6.56 and parameters: {'n_estimators': 255, 'max_depth': 1, 'min_child_samples': 8, 'learning_rate': 0.06999999999999999, 'subsample': 0.88, 'colsample_bytree': 0.8999999999999999}. Best is trial 0 with value: 6.35.
[I 2023-08-24 10:27:19,194] Trial 5 finished with value: 6.56 and parameters: {'n_estimators': 235, 'max_depth': 9, 'min_child_samples': 10, 'learning_rate': 0.09, 'subsample': 0.7, 'colsample_bytree': 0.8999999999999999}. Best is trial 0 with value: 6.35.
[I 2023-08-24 10:27:19,502] Trial 6 finished with value: 6.48 and parameters: {'n_estimators': 112, 'max_depth': 10, 'min_child_samples': 10, 'learning_rate': 0.02, 'subsample': 0.87, 'colsample_bytree': 1.0}. Best is trial 0 with value: 6.35.
[I 2023-08-24 10:27:20,048] Trial 7 finished with value: 6.53 and parameters: {'n_estimators': 369, 'max_depth': 6, 'min_child_samples': 6, 'learning_rate': 0.06999999999999999, 'subsample': 0.62, 'colsample_bytree': 0.7999999999999999}. Best is trial 0 with value: 6.35.
[I 2023-08-24 10:27:20,337] Trial 8 finished with value: 6.43 and parameters: {'n_estimators': 154, 'max_depth': 8, 'min_child_samples': 6, 'learning_rate': 0.06999999999999999, 'subsample': 0.8500000000000001, 'colsample_bytree': 0.7}. Best is trial 0 with value: 6.35.
[I 2023-08-24 10:27:20,488] Trial 9 finished with value: 6.4 and parameters: {'n_estimators': 119, 'max_depth': 5, 'min_child_samples': 9, 'learning_rate': 0.09, 'subsample': 0.76, 'colsample_bytree': 0.8999999999999999}. Best is trial 0 with value: 6.35.
[I 2023-08-24 10:27:20,624] Trial 10 finished with value: 6.49 and parameters: {'n_estimators': 193, 'max_depth': 2, 'min_child_samples': 9, 'learning_rate': 0.04, 'subsample': 0.51, 'colsample_bytree': 0.7999999999999999}. Best is trial 0 with value: 6.35.
[I 2023-08-24 10:27:20,825] Trial 11 finished with value: 6.29 and parameters: {'n_estimators': 163, 'max_depth': 4, 'min_child_samples': 9, 'learning_rate': 0.04, 'subsample': 0.78, 'colsample_bytree': 0.8999999999999999}. Best is trial 11 with value: 6.29.
[I 2023-08-24 10:27:20,983] Trial 12 finished with value: 6.37 and parameters: {'n_estimators': 164, 'max_depth': 3, 'min_child_samples': 9, 'learning_rate': 0.04, 'subsample': 0.79, 'colsample_bytree': 0.7999999999999999}. Best is trial 11 with value: 6.29.
[I 2023-08-24 10:27:21,266] Trial 13 finished with value: 6.28 and parameters: {'n_estimators': 300, 'max_depth': 4, 'min_child_samples': 8, 'learning_rate': 0.04, 'subsample': 0.78, 'colsample_bytree': 0.8999999999999999}. Best is trial 13 with value: 6.28.
[I 2023-08-24 10:27:21,844] Trial 14 finished with value: 6.49 and parameters: {'n_estimators': 301, 'max_depth': 7, 'min_child_samples': 9, 'learning_rate': 0.05, 'subsample': 0.76, 'colsample_bytree': 1.0}. Best is trial 13 with value: 6.28.
[I 2023-08-24 10:27:22,233] Trial 15 finished with value: 6.51 and parameters: {'n_estimators': 314, 'max_depth': 4, 'min_child_samples': 7, 'learning_rate': 0.01, 'subsample': 0.79, 'colsample_bytree': 0.8999999999999999}. Best is trial 13 with value: 6.28.
[I 2023-08-24 10:27:22,484] Trial 16 finished with value: 6.29 and parameters: {'n_estimators': 306, 'max_depth': 3, 'min_child_samples': 10, 'learning_rate': 0.05, 'subsample': 0.73, 'colsample_bytree': 0.8999999999999999}. Best is trial 13 with value: 6.28.
[I 2023-08-24 10:27:22,910] Trial 17 finished with value: 6.37 and parameters: {'n_estimators': 203, 'max_depth': 6, 'min_child_samples': 9, 'learning_rate': 0.03, 'subsample': 0.65, 'colsample_bytree': 1.0}. Best is trial 13 with value: 6.28.
[I 2023-08-24 10:27:23,040] Trial 18 finished with value: 6.58 and parameters: {'n_estimators': 275, 'max_depth': 1, 'min_child_samples': 8, 'learning_rate': 0.060000000000000005, 'subsample': 0.81, 'colsample_bytree': 0.8999999999999999}. Best is trial 13 with value: 6.28.
[I 2023-08-24 10:27:23,274] Trial 19 finished with value: 6.3 and parameters: {'n_estimators': 336, 'max_depth': 3, 'min_child_samples': 7, 'learning_rate': 0.03, 'subsample': 0.73, 'colsample_bytree': 0.7999999999999999}. Best is trial 13 with value: 6.28.
[I 2023-08-24 10:27:23,538] Trial 20 finished with value: 6.68 and parameters: {'n_estimators': 207, 'max_depth': 4, 'min_child_samples': 8, 'learning_rate': 0.01, 'subsample': 0.8200000000000001, 'colsample_bytree': 1.0}. Best is trial 13 with value: 6.28.
[I 2023-08-24 10:27:23,764] Trial 21 finished with value: 6.29 and parameters: {'n_estimators': 290, 'max_depth': 3, 'min_child_samples': 10, 'learning_rate': 0.05, 'subsample': 0.73, 'colsample_bytree': 0.8999999999999999}. Best is trial 13 with value: 6.28.
[I 2023-08-24 10:27:23,946] Trial 22 finished with value: 6.31 and parameters: {'n_estimators': 332, 'max_depth': 2, 'min_child_samples': 10, 'learning_rate': 0.060000000000000005, 'subsample': 0.75, 'colsample_bytree': 0.8999999999999999}. Best is trial 13 with value: 6.28.
[I 2023-08-24 10:27:24,362] Trial 23 finished with value: 6.4 and parameters: {'n_estimators': 342, 'max_depth': 5, 'min_child_samples': 10, 'learning_rate': 0.05, 'subsample': 0.6799999999999999, 'colsample_bytree': 0.8999999999999999}. Best is trial 13 with value: 6.28.
[I 2023-08-24 10:27:24,518] Trial 24 finished with value: 6.5 and parameters: {'n_estimators': 249, 'max_depth': 2, 'min_child_samples': 9, 'learning_rate': 0.03, 'subsample': 0.72, 'colsample_bytree': 0.8999999999999999}. Best is trial 13 with value: 6.28.
[I 2023-08-24 10:27:24,791] Trial 25 finished with value: 6.27 and parameters: {'n_estimators': 278, 'max_depth': 4, 'min_child_samples': 9, 'learning_rate': 0.04, 'subsample': 0.77, 'colsample_bytree': 0.7999999999999999}. Best is trial 25 with value: 6.27.
[I 2023-08-24 10:27:25,055] Trial 26 finished with value: 6.26 and parameters: {'n_estimators': 269, 'max_depth': 4, 'min_child_samples': 8, 'learning_rate': 0.04, 'subsample': 0.77, 'colsample_bytree': 0.7999999999999999}. Best is trial 26 with value: 6.26.
[I 2023-08-24 10:27:25,794] Trial 27 finished with value: 6.33 and parameters: {'n_estimators': 396, 'max_depth': 7, 'min_child_samples': 8, 'learning_rate': 0.02, 'subsample': 0.8300000000000001, 'colsample_bytree': 0.7999999999999999}. Best is trial 26 with value: 6.26.
[I 2023-08-24 10:27:26,142] Trial 28 finished with value: 6.3 and parameters: {'n_estimators': 267, 'max_depth': 5, 'min_child_samples': 8, 'learning_rate': 0.03, 'subsample': 0.8, 'colsample_bytree': 0.7999999999999999}. Best is trial 26 with value: 6.26.
[I 2023-08-24 10:27:26,538] Trial 29 finished with value: 6.3 and parameters: {'n_estimators': 226, 'max_depth': 6, 'min_child_samples': 7, 'learning_rate': 0.04, 'subsample': 0.9, 'colsample_bytree': 0.7}. Best is trial 26 with value: 6.26.
[I 2023-08-24 10:27:26,806] Trial 30 finished with value: 6.34 and parameters: {'n_estimators': 281, 'max_depth': 4, 'min_child_samples': 8, 'learning_rate': 0.060000000000000005, 'subsample': 0.8500000000000001, 'colsample_bytree': 0.7999999999999999}. Best is trial 26 with value: 6.26.
[I 2023-08-24 10:27:27,010] Trial 31 finished with value: 6.31 and parameters: {'n_estimators': 156, 'max_depth': 4, 'min_child_samples': 9, 'learning_rate': 0.04, 'subsample': 0.77, 'colsample_bytree': 0.7999999999999999}. Best is trial 26 with value: 6.26.
[I 2023-08-24 10:27:27,342] Trial 32 finished with value: 6.32 and parameters: {'n_estimators': 263, 'max_depth': 5, 'min_child_samples': 9, 'learning_rate': 0.03, 'subsample': 0.78, 'colsample_bytree': 0.7999999999999999}. Best is trial 26 with value: 6.26.
[I 2023-08-24 10:27:27,543] Trial 33 finished with value: 6.28 and parameters: {'n_estimators': 178, 'max_depth': 4, 'min_child_samples': 8, 'learning_rate': 0.04, 'subsample': 0.69, 'colsample_bytree': 0.7}. Best is trial 26 with value: 6.26.
[I 2023-08-24 10:27:27,723] Trial 34 finished with value: 6.31 and parameters: {'n_estimators': 224, 'max_depth': 3, 'min_child_samples': 8, 'learning_rate': 0.04, 'subsample': 0.64, 'colsample_bytree': 0.7}. Best is trial 26 with value: 6.26.
[I 2023-08-24 10:27:28,013] Trial 35 finished with value: 6.28 and parameters: {'n_estimators': 323, 'max_depth': 4, 'min_child_samples': 7, 'learning_rate': 0.05, 'subsample': 0.7, 'colsample_bytree': 0.7}. Best is trial 26 with value: 6.26.
[I 2023-08-24 10:27:28,203] Trial 36 finished with value: 6.52 and parameters: {'n_estimators': 356, 'max_depth': 2, 'min_child_samples': 8, 'learning_rate': 0.02, 'subsample': 0.59, 'colsample_bytree': 0.7}. Best is trial 26 with value: 6.26.
[I 2023-08-24 10:27:28,572] Trial 37 finished with value: 6.29 and parameters: {'n_estimators': 294, 'max_depth': 5, 'min_child_samples': 8, 'learning_rate': 0.03, 'subsample': 0.6799999999999999, 'colsample_bytree': 0.7}. Best is trial 26 with value: 6.26.
[I 2023-08-24 10:27:28,770] Trial 38 finished with value: 6.32 and parameters: {'n_estimators': 246, 'max_depth': 3, 'min_child_samples': 7, 'learning_rate': 0.060000000000000005, 'subsample': 0.66, 'colsample_bytree': 0.7999999999999999}. Best is trial 26 with value: 6.26.
[I 2023-08-24 10:27:29,079] Trial 39 finished with value: 6.5 and parameters: {'n_estimators': 182, 'max_depth': 6, 'min_child_samples': 8, 'learning_rate': 0.09999999999999999, 'subsample': 0.75, 'colsample_bytree': 0.7}. Best is trial 26 with value: 6.26.
[I 2023-08-24 10:27:29,248] Trial 40 finished with value: 6.56 and parameters: {'n_estimators': 131, 'max_depth': 4, 'min_child_samples': 8, 'learning_rate': 0.02, 'subsample': 0.54, 'colsample_bytree': 0.7}. Best is trial 26 with value: 6.26.
[I 2023-08-24 10:27:29,531] Trial 41 finished with value: 6.27 and parameters: {'n_estimators': 312, 'max_depth': 4, 'min_child_samples': 7, 'learning_rate': 0.05, 'subsample': 0.7, 'colsample_bytree': 0.7}. Best is trial 26 with value: 6.26.
[I 2023-08-24 10:27:29,798] Trial 42 finished with value: 6.29 and parameters: {'n_estimators': 275, 'max_depth': 4, 'min_child_samples': 6, 'learning_rate': 0.04, 'subsample': 0.71, 'colsample_bytree': 0.7}. Best is trial 26 with value: 6.26.
[I 2023-08-24 10:27:30,176] Trial 43 finished with value: 6.32 and parameters: {'n_estimators': 312, 'max_depth': 5, 'min_child_samples': 7, 'learning_rate': 0.05, 'subsample': 0.69, 'colsample_bytree': 0.7}. Best is trial 26 with value: 6.26.
[I 2023-08-24 10:27:30,397] Trial 44 finished with value: 6.31 and parameters: {'n_estimators': 289, 'max_depth': 3, 'min_child_samples': 7, 'learning_rate': 0.04, 'subsample': 0.61, 'colsample_bytree': 0.7999999999999999}. Best is trial 26 with value: 6.26.
[I 2023-08-24 10:27:30,707] Trial 45 finished with value: 6.34 and parameters: {'n_estimators': 233, 'max_depth': 5, 'min_child_samples': 8, 'learning_rate': 0.05, 'subsample': 0.74, 'colsample_bytree': 0.7999999999999999}. Best is trial 26 with value: 6.26.
[I 2023-08-24 10:27:31,166] Trial 46 finished with value: 6.53 and parameters: {'n_estimators': 254, 'max_depth': 7, 'min_child_samples': 6, 'learning_rate': 0.06999999999999999, 'subsample': 0.67, 'colsample_bytree': 0.7}. Best is trial 26 with value: 6.26.
[I 2023-08-24 10:27:31,351] Trial 47 finished with value: 6.34 and parameters: {'n_estimators': 140, 'max_depth': 4, 'min_child_samples': 7, 'learning_rate': 0.04, 'subsample': 0.7, 'colsample_bytree': 0.7999999999999999}. Best is trial 26 with value: 6.26.
[I 2023-08-24 10:27:31,579] Trial 48 finished with value: 6.26 and parameters: {'n_estimators': 321, 'max_depth': 3, 'min_child_samples': 9, 'learning_rate': 0.060000000000000005, 'subsample': 0.8400000000000001, 'colsample_bytree': 0.7}. Best is trial 26 with value: 6.26.
[I 2023-08-24 10:27:31,768] Trial 49 finished with value: 6.28 and parameters: {'n_estimators': 349, 'max_depth': 2, 'min_child_samples': 9, 'learning_rate': 0.08, 'subsample': 0.87, 'colsample_bytree': 0.7999999999999999}. Best is trial 26 with value: 6.26.

Лучшие параметры: {'n_estimators': 269, 'max_depth': 4, 'min_child_samples': 8, 'learning_rate': 0.04, 'subsample': 0.77, 'colsample_bytree': 0.7999999999999999}
Лучшее значение метрики: 6.26

<timed exec>:57: ExperimentalWarning:

plot_param_importances is experimental (supported from v2.2.0). The interface can change in the future.

<timed exec>:73: ExperimentalWarning:

plot_optimization_history is experimental (supported from v2.2.0). The interface can change in the future.

Wall time: 17.2 s


# Создаём график контура
fig = optuna.visualization.plot_contour(study, params=["max_depth", "n_estimators"])
fig.show()


# Создаём график контура
fig = optuna.visualization.plot_contour(study, params=["learning_rate", "n_estimators"])
fig.show()


# Создаём график контура
fig = optuna.visualization.plot_contour(study, params=["subsample", "n_estimators"])
fig.show()


%%time

def objective(trial):
    # Определяем параметры для подбора
    n_estimators = trial.suggest_int('n_estimators', 150, 350)
    max_depth = trial.suggest_int('max_depth', 2, 6)
    min_child_samples = trial.suggest_int('min_child_samples', 6, 8)
    learning_rate = trial.suggest_float('learning_rate', 0.01, 0.1, step=0.01)
    subsample = trial.suggest_float('subsample', 0.5, 0.9, step=0.01)
    colsample_bytree = trial.suggest_float('colsample_bytree', 0.7, 1, step=0.1)
    
    # Создаем модель с определенными параметрами
    model = LGBMRegressor(n_estimators=n_estimators,
                          max_depth=max_depth,
                          min_child_samples=min_child_samples,
                          learning_rate=learning_rate,
                          subsample=subsample,
                          colsample_bytree=colsample_bytree,
                          random_state=RANDOM_STATE)
    
    # Оцениваем качество модели с помощью кросс-валидации
    score = round(abs(cross_val_score(model,
                                      features_train,
                                      target_train,
                                      cv=kf,
                                      scoring='neg_mean_absolute_error', n_jobs=-1)).mean(), 2)
    
    # Возвращаем значения метрики
    return score.mean()

# Создаем объект Study и запускаем оптимизацию
study = optuna.create_study(direction='minimize')
study.optimize(objective, n_trials=200)

print('Лучшие параметры:', study.best_params)
print('Лучшее значение метрики:', study.best_value)

# Создаем датафрейм с значениями гиперпараметров и метрикой
results = pd.DataFrame({'n_estimators': [],
                       'max_depth': [],
                       'min_child_samples': [],
                       'learning_rate': [],
                       'subsample': [],
                       'colsample_bytree': [],
                       'score': []})
# Заполняем данными из best_trials
for trial in study.best_trials:
    params = trial.params
    params['score'] = trial.value
    results = results.append(params, ignore_index=True)

# Сортируем датафрейм по убыванию значения метрики
results = results.sort_values(by='score', ascending=False)

# Выводим 10 лучших моделей
display(results.head(10))

# Создаем график важности параметров
optuna.visualization.matplotlib.plot_param_importances(study)
plt.xlabel('Важность гиперпараметров',
           fontsize=12,
           color='DarkSlateGray')
plt.ylabel('Гиперпараметры',
           fontsize=12,
           color='DarkSlateGray')
plt.suptitle('Важность гиперпараметров',
             fontsize=15,
             color='DarkSlateGray')
plt.minorticks_on()
plt.grid(which='minor',
         linestyle=':')
plt.grid(True)

# Создаем график истории оптимизации
optuna.visualization.matplotlib.plot_optimization_history(study)
plt.tight_layout()
plt.xlabel('Итерации',
           fontsize=12,
           color='DarkSlateGray')
plt.ylabel('Целевая метрика',
           fontsize=12,
           color='DarkSlateGray')
plt.suptitle('История оптимизации \n \n',
             fontsize=15,
             color='DarkSlateGray')
plt.minorticks_on()
plt.grid(which='minor',
         linestyle=':')
plt.grid(True)
plt.show()

# Сохранение модели с наилучшими параметрами
lgbm = LGBMRegressor(n_estimators=study.best_params['n_estimators'],
                     max_depth=study.best_params['max_depth'],
                     min_child_samples=study.best_params['min_child_samples'],
                     learning_rate=study.best_params['learning_rate'],
                     subsample=study.best_params['subsample'],
                     colsample_bytree=study.best_params['colsample_bytree'],
                     random_state=RANDOM_STATE)

[I 2023-08-24 10:27:34,602] A new study created in memory with name: no-name-3796d158-4163-482b-ad38-0999a9d62315
[I 2023-08-24 10:27:34,746] Trial 0 finished with value: 6.3 and parameters: {'n_estimators': 290, 'max_depth': 2, 'min_child_samples': 6, 'learning_rate': 0.09999999999999999, 'subsample': 0.8, 'colsample_bytree': 1.0}. Best is trial 0 with value: 6.3.
[I 2023-08-24 10:27:34,934] Trial 1 finished with value: 6.37 and parameters: {'n_estimators': 265, 'max_depth': 3, 'min_child_samples': 6, 'learning_rate': 0.08, 'subsample': 0.78, 'colsample_bytree': 1.0}. Best is trial 0 with value: 6.3.
[I 2023-08-24 10:27:35,130] Trial 2 finished with value: 6.39 and parameters: {'n_estimators': 197, 'max_depth': 4, 'min_child_samples': 8, 'learning_rate': 0.09, 'subsample': 0.81, 'colsample_bytree': 1.0}. Best is trial 0 with value: 6.3.
[I 2023-08-24 10:27:35,224] Trial 3 finished with value: 6.31 and parameters: {'n_estimators': 190, 'max_depth': 2, 'min_child_samples': 8, 'learning_rate': 0.09999999999999999, 'subsample': 0.55, 'colsample_bytree': 0.7}. Best is trial 0 with value: 6.3.
[I 2023-08-24 10:27:35,418] Trial 4 finished with value: 6.29 and parameters: {'n_estimators': 151, 'max_depth': 4, 'min_child_samples': 8, 'learning_rate': 0.08, 'subsample': 0.72, 'colsample_bytree': 1.0}. Best is trial 4 with value: 6.29.
[I 2023-08-24 10:27:35,519] Trial 5 finished with value: 6.36 and parameters: {'n_estimators': 195, 'max_depth': 2, 'min_child_samples': 6, 'learning_rate': 0.060000000000000005, 'subsample': 0.66, 'colsample_bytree': 0.7}. Best is trial 4 with value: 6.29.
[I 2023-08-24 10:27:35,663] Trial 6 finished with value: 6.3 and parameters: {'n_estimators': 337, 'max_depth': 2, 'min_child_samples': 8, 'learning_rate': 0.060000000000000005, 'subsample': 0.72, 'colsample_bytree': 0.7999999999999999}. Best is trial 4 with value: 6.29.
[I 2023-08-24 10:27:35,799] Trial 7 finished with value: 6.29 and parameters: {'n_estimators': 295, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.78, 'colsample_bytree': 0.7999999999999999}. Best is trial 4 with value: 6.29.
[I 2023-08-24 10:27:36,345] Trial 8 finished with value: 6.34 and parameters: {'n_estimators': 340, 'max_depth': 6, 'min_child_samples': 7, 'learning_rate': 0.02, 'subsample': 0.8200000000000001, 'colsample_bytree': 0.7}. Best is trial 4 with value: 6.29.
[I 2023-08-24 10:27:36,628] Trial 9 finished with value: 6.46 and parameters: {'n_estimators': 194, 'max_depth': 6, 'min_child_samples': 8, 'learning_rate': 0.09999999999999999, 'subsample': 0.72, 'colsample_bytree': 0.7999999999999999}. Best is trial 4 with value: 6.29.
[I 2023-08-24 10:27:36,877] Trial 10 finished with value: 6.36 and parameters: {'n_estimators': 152, 'max_depth': 5, 'min_child_samples': 7, 'learning_rate': 0.03, 'subsample': 0.9, 'colsample_bytree': 0.8999999999999999}. Best is trial 4 with value: 6.29.
[I 2023-08-24 10:27:37,179] Trial 11 finished with value: 6.42 and parameters: {'n_estimators': 298, 'max_depth': 4, 'min_child_samples': 7, 'learning_rate': 0.08, 'subsample': 0.64, 'colsample_bytree': 0.8999999999999999}. Best is trial 4 with value: 6.29.
[I 2023-08-24 10:27:37,372] Trial 12 finished with value: 6.3 and parameters: {'n_estimators': 238, 'max_depth': 3, 'min_child_samples': 7, 'learning_rate': 0.06999999999999999, 'subsample': 0.58, 'colsample_bytree': 0.8999999999999999}. Best is trial 4 with value: 6.29.
[I 2023-08-24 10:27:37,684] Trial 13 finished with value: 6.41 and parameters: {'n_estimators': 240, 'max_depth': 5, 'min_child_samples': 7, 'learning_rate': 0.08, 'subsample': 0.9, 'colsample_bytree': 0.7999999999999999}. Best is trial 4 with value: 6.29.
[I 2023-08-24 10:27:37,841] Trial 14 finished with value: 6.35 and parameters: {'n_estimators': 152, 'max_depth': 3, 'min_child_samples': 8, 'learning_rate': 0.05, 'subsample': 0.73, 'colsample_bytree': 0.8999999999999999}. Best is trial 4 with value: 6.29.
[I 2023-08-24 10:27:38,229] Trial 15 finished with value: 6.31 and parameters: {'n_estimators': 311, 'max_depth': 5, 'min_child_samples': 7, 'learning_rate': 0.04, 'subsample': 0.5, 'colsample_bytree': 0.7999999999999999}. Best is trial 4 with value: 6.29.
[I 2023-08-24 10:27:38,438] Trial 16 finished with value: 6.39 and parameters: {'n_estimators': 262, 'max_depth': 3, 'min_child_samples': 8, 'learning_rate': 0.09, 'subsample': 0.67, 'colsample_bytree': 1.0}. Best is trial 4 with value: 6.29.
[I 2023-08-24 10:27:38,673] Trial 17 finished with value: 6.35 and parameters: {'n_estimators': 217, 'max_depth': 4, 'min_child_samples': 6, 'learning_rate': 0.06999999999999999, 'subsample': 0.76, 'colsample_bytree': 0.8999999999999999}. Best is trial 4 with value: 6.29.
[I 2023-08-24 10:27:38,987] Trial 18 finished with value: 6.54 and parameters: {'n_estimators': 283, 'max_depth': 4, 'min_child_samples': 7, 'learning_rate': 0.01, 'subsample': 0.86, 'colsample_bytree': 0.7999999999999999}. Best is trial 4 with value: 6.29.
[I 2023-08-24 10:27:39,201] Trial 19 finished with value: 6.37 and parameters: {'n_estimators': 313, 'max_depth': 3, 'min_child_samples': 8, 'learning_rate': 0.09, 'subsample': 0.62, 'colsample_bytree': 0.7}. Best is trial 4 with value: 6.29.
[I 2023-08-24 10:27:39,455] Trial 20 finished with value: 6.36 and parameters: {'n_estimators': 170, 'max_depth': 5, 'min_child_samples': 7, 'learning_rate': 0.06999999999999999, 'subsample': 0.69, 'colsample_bytree': 1.0}. Best is trial 4 with value: 6.29.
[I 2023-08-24 10:27:39,625] Trial 21 finished with value: 6.31 and parameters: {'n_estimators': 279, 'max_depth': 2, 'min_child_samples': 6, 'learning_rate': 0.09999999999999999, 'subsample': 0.77, 'colsample_bytree': 1.0}. Best is trial 4 with value: 6.29.
[I 2023-08-24 10:27:39,804] Trial 22 finished with value: 6.31 and parameters: {'n_estimators': 312, 'max_depth': 2, 'min_child_samples': 6, 'learning_rate': 0.09999999999999999, 'subsample': 0.8300000000000001, 'colsample_bytree': 1.0}. Best is trial 4 with value: 6.29.
[I 2023-08-24 10:27:39,976] Trial 23 finished with value: 6.32 and parameters: {'n_estimators': 291, 'max_depth': 2, 'min_child_samples': 6, 'learning_rate': 0.09, 'subsample': 0.75, 'colsample_bytree': 0.8999999999999999}. Best is trial 4 with value: 6.29.
[I 2023-08-24 10:27:40,189] Trial 24 finished with value: 6.37 and parameters: {'n_estimators': 261, 'max_depth': 3, 'min_child_samples': 6, 'learning_rate': 0.08, 'subsample': 0.86, 'colsample_bytree': 1.0}. Best is trial 4 with value: 6.29.
[I 2023-08-24 10:27:40,335] Trial 25 finished with value: 6.29 and parameters: {'n_estimators': 223, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.8, 'colsample_bytree': 0.8999999999999999}. Best is trial 4 with value: 6.29.
[I 2023-08-24 10:27:40,576] Trial 26 finished with value: 6.41 and parameters: {'n_estimators': 221, 'max_depth': 4, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.86, 'colsample_bytree': 0.7999999999999999}. Best is trial 4 with value: 6.29.
[I 2023-08-24 10:27:40,740] Trial 27 finished with value: 6.32 and parameters: {'n_estimators': 170, 'max_depth': 3, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.79, 'colsample_bytree': 0.8999999999999999}. Best is trial 4 with value: 6.29.
[I 2023-08-24 10:27:40,880] Trial 28 finished with value: 6.37 and parameters: {'n_estimators': 224, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.05, 'subsample': 0.69, 'colsample_bytree': 0.7999999999999999}. Best is trial 4 with value: 6.29.
[I 2023-08-24 10:27:41,017] Trial 29 finished with value: 6.33 and parameters: {'n_estimators': 176, 'max_depth': 2, 'min_child_samples': 8, 'learning_rate': 0.09999999999999999, 'subsample': 0.73, 'colsample_bytree': 0.8999999999999999}. Best is trial 4 with value: 6.29.
[I 2023-08-24 10:27:41,433] Trial 30 finished with value: 6.47 and parameters: {'n_estimators': 242, 'max_depth': 6, 'min_child_samples': 7, 'learning_rate': 0.08, 'subsample': 0.8400000000000001, 'colsample_bytree': 0.8999999999999999}. Best is trial 4 with value: 6.29.
[I 2023-08-24 10:27:41,600] Trial 31 finished with value: 6.3 and parameters: {'n_estimators': 272, 'max_depth': 2, 'min_child_samples': 6, 'learning_rate': 0.09999999999999999, 'subsample': 0.79, 'colsample_bytree': 1.0}. Best is trial 4 with value: 6.29.
[I 2023-08-24 10:27:41,851] Trial 32 finished with value: 6.48 and parameters: {'n_estimators': 325, 'max_depth': 3, 'min_child_samples': 6, 'learning_rate': 0.09999999999999999, 'subsample': 0.8, 'colsample_bytree': 1.0}. Best is trial 4 with value: 6.29.
[I 2023-08-24 10:27:42,039] Trial 33 finished with value: 6.32 and parameters: {'n_estimators': 296, 'max_depth': 2, 'min_child_samples': 8, 'learning_rate': 0.09, 'subsample': 0.75, 'colsample_bytree': 1.0}. Best is trial 4 with value: 6.29.
[I 2023-08-24 10:27:42,248] Trial 34 finished with value: 6.35 and parameters: {'n_estimators': 255, 'max_depth': 3, 'min_child_samples': 8, 'learning_rate': 0.08, 'subsample': 0.78, 'colsample_bytree': 1.0}. Best is trial 4 with value: 6.29.
[I 2023-08-24 10:27:42,428] Trial 35 finished with value: 6.3 and parameters: {'n_estimators': 327, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.81, 'colsample_bytree': 1.0}. Best is trial 4 with value: 6.29.
[I 2023-08-24 10:27:42,572] Trial 36 finished with value: 6.3 and parameters: {'n_estimators': 204, 'max_depth': 2, 'min_child_samples': 6, 'learning_rate': 0.09999999999999999, 'subsample': 0.71, 'colsample_bytree': 0.7}. Best is trial 4 with value: 6.29.
[I 2023-08-24 10:27:42,705] Trial 37 finished with value: 6.36 and parameters: {'n_estimators': 182, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.06999999999999999, 'subsample': 0.75, 'colsample_bytree': 0.7999999999999999}. Best is trial 4 with value: 6.29.
[I 2023-08-24 10:27:42,895] Trial 38 finished with value: 6.31 and parameters: {'n_estimators': 207, 'max_depth': 3, 'min_child_samples': 8, 'learning_rate': 0.060000000000000005, 'subsample': 0.8300000000000001, 'colsample_bytree': 1.0}. Best is trial 4 with value: 6.29.
[I 2023-08-24 10:27:43,071] Trial 39 finished with value: 6.27 and parameters: {'n_estimators': 304, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.77, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:43,264] Trial 40 finished with value: 6.29 and parameters: {'n_estimators': 350, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.67, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:43,449] Trial 41 finished with value: 6.29 and parameters: {'n_estimators': 344, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.67, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:43,640] Trial 42 finished with value: 6.28 and parameters: {'n_estimators': 330, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.62, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:43,830] Trial 43 finished with value: 6.31 and parameters: {'n_estimators': 325, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.58, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:44,058] Trial 44 finished with value: 6.33 and parameters: {'n_estimators': 307, 'max_depth': 3, 'min_child_samples': 7, 'learning_rate': 0.08, 'subsample': 0.64, 'colsample_bytree': 0.7999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:44,233] Trial 45 finished with value: 6.28 and parameters: {'n_estimators': 305, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.6, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:44,548] Trial 46 finished with value: 6.48 and parameters: {'n_estimators': 331, 'max_depth': 4, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.6, 'colsample_bytree': 0.7999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:44,785] Trial 47 finished with value: 6.34 and parameters: {'n_estimators': 302, 'max_depth': 3, 'min_child_samples': 7, 'learning_rate': 0.08, 'subsample': 0.54, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:45,304] Trial 48 finished with value: 6.62 and parameters: {'n_estimators': 317, 'max_depth': 6, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.54, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:45,660] Trial 49 finished with value: 6.51 and parameters: {'n_estimators': 282, 'max_depth': 5, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.64, 'colsample_bytree': 0.7999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:45,848] Trial 50 finished with value: 6.42 and parameters: {'n_estimators': 320, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.03, 'subsample': 0.58, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:46,044] Trial 51 finished with value: 6.28 and parameters: {'n_estimators': 335, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.61, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:46,234] Trial 52 finished with value: 6.28 and parameters: {'n_estimators': 336, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.6, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:46,425] Trial 53 finished with value: 6.29 and parameters: {'n_estimators': 339, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.61, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:46,617] Trial 54 finished with value: 6.31 and parameters: {'n_estimators': 331, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.56, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:46,810] Trial 55 finished with value: 6.29 and parameters: {'n_estimators': 350, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.56, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:46,993] Trial 56 finished with value: 6.28 and parameters: {'n_estimators': 305, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.6, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:47,175] Trial 57 finished with value: 6.27 and parameters: {'n_estimators': 302, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.51, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:47,367] Trial 58 finished with value: 6.3 and parameters: {'n_estimators': 336, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.5, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:47,539] Trial 59 finished with value: 6.28 and parameters: {'n_estimators': 317, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.52, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:47,779] Trial 60 finished with value: 6.39 and parameters: {'n_estimators': 287, 'max_depth': 3, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.62, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:47,956] Trial 61 finished with value: 6.27 and parameters: {'n_estimators': 302, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.6, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:48,144] Trial 62 finished with value: 6.27 and parameters: {'n_estimators': 303, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.65, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:48,311] Trial 63 finished with value: 6.27 and parameters: {'n_estimators': 271, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.65, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:48,488] Trial 64 finished with value: 6.3 and parameters: {'n_estimators': 275, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.63, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:48,666] Trial 65 finished with value: 6.27 and parameters: {'n_estimators': 292, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.65, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:48,856] Trial 66 finished with value: 6.27 and parameters: {'n_estimators': 290, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.66, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:49,037] Trial 67 finished with value: 6.3 and parameters: {'n_estimators': 291, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.66, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:49,201] Trial 68 finished with value: 6.27 and parameters: {'n_estimators': 268, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.65, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:49,428] Trial 69 finished with value: 6.4 and parameters: {'n_estimators': 298, 'max_depth': 3, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.69, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:49,645] Trial 70 finished with value: 6.28 and parameters: {'n_estimators': 287, 'max_depth': 3, 'min_child_samples': 7, 'learning_rate': 0.04, 'subsample': 0.6799999999999999, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:49,825] Trial 71 finished with value: 6.27 and parameters: {'n_estimators': 269, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.65, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:49,997] Trial 72 finished with value: 6.27 and parameters: {'n_estimators': 266, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.66, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:50,161] Trial 73 finished with value: 6.31 and parameters: {'n_estimators': 251, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.71, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:50,338] Trial 74 finished with value: 6.86 and parameters: {'n_estimators': 279, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.01, 'subsample': 0.65, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:50,515] Trial 75 finished with value: 6.27 and parameters: {'n_estimators': 296, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.7, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:50,685] Trial 76 finished with value: 6.3 and parameters: {'n_estimators': 276, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.65, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:50,869] Trial 77 finished with value: 6.28 and parameters: {'n_estimators': 310, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.6799999999999999, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:51,039] Trial 78 finished with value: 6.27 and parameters: {'n_estimators': 261, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.63, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:51,220] Trial 79 finished with value: 6.3 and parameters: {'n_estimators': 291, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.73, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:51,447] Trial 80 finished with value: 6.4 and parameters: {'n_estimators': 301, 'max_depth': 3, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.58, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:51,609] Trial 81 finished with value: 6.27 and parameters: {'n_estimators': 269, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.64, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:51,788] Trial 82 finished with value: 6.27 and parameters: {'n_estimators': 284, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.66, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:51,966] Trial 83 finished with value: 6.27 and parameters: {'n_estimators': 259, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.65, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:52,131] Trial 84 finished with value: 6.3 and parameters: {'n_estimators': 267, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.6799999999999999, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:52,296] Trial 85 finished with value: 6.27 and parameters: {'n_estimators': 249, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.63, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:52,473] Trial 86 finished with value: 6.27 and parameters: {'n_estimators': 272, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.67, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:52,651] Trial 87 finished with value: 6.3 and parameters: {'n_estimators': 293, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.7, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:52,816] Trial 88 finished with value: 6.28 and parameters: {'n_estimators': 233, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.52, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:53,003] Trial 89 finished with value: 6.31 and parameters: {'n_estimators': 301, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.65, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:53,229] Trial 90 finished with value: 6.33 and parameters: {'n_estimators': 284, 'max_depth': 3, 'min_child_samples': 7, 'learning_rate': 0.08, 'subsample': 0.61, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:53,405] Trial 91 finished with value: 6.27 and parameters: {'n_estimators': 266, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.66, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:53,571] Trial 92 finished with value: 6.28 and parameters: {'n_estimators': 254, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.66, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:53,741] Trial 93 finished with value: 6.27 and parameters: {'n_estimators': 278, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.63, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:53,924] Trial 94 finished with value: 6.27 and parameters: {'n_estimators': 271, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.6799999999999999, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:54,121] Trial 95 finished with value: 6.28 and parameters: {'n_estimators': 309, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.69, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:54,305] Trial 96 finished with value: 6.28 and parameters: {'n_estimators': 315, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.64, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:54,497] Trial 97 finished with value: 6.3 and parameters: {'n_estimators': 322, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.59, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:54,671] Trial 98 finished with value: 6.29 and parameters: {'n_estimators': 288, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.67, 'colsample_bytree': 0.7999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:54,859] Trial 99 finished with value: 6.3 and parameters: {'n_estimators': 297, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.060000000000000005, 'subsample': 0.71, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:55,035] Trial 100 finished with value: 6.3 and parameters: {'n_estimators': 263, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.62, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:55,216] Trial 101 finished with value: 6.27 and parameters: {'n_estimators': 294, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.89, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:55,382] Trial 102 finished with value: 6.28 and parameters: {'n_estimators': 245, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.74, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:55,570] Trial 103 finished with value: 6.28 and parameters: {'n_estimators': 306, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.66, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:55,748] Trial 104 finished with value: 6.27 and parameters: {'n_estimators': 282, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.7, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:56,199] Trial 105 finished with value: 6.31 and parameters: {'n_estimators': 299, 'max_depth': 5, 'min_child_samples': 7, 'learning_rate': 0.02, 'subsample': 0.67, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:56,388] Trial 106 finished with value: 6.32 and parameters: {'n_estimators': 312, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.05, 'subsample': 0.77, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:56,568] Trial 107 finished with value: 6.27 and parameters: {'n_estimators': 303, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.65, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:56,732] Trial 108 finished with value: 6.31 and parameters: {'n_estimators': 256, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.72, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:56,921] Trial 109 finished with value: 6.27 and parameters: {'n_estimators': 280, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.64, 'colsample_bytree': 0.8999999999999999}. Best is trial 39 with value: 6.27.
[I 2023-08-24 10:27:57,089] Trial 110 finished with value: 6.26 and parameters: {'n_estimators': 275, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.7, 'colsample_bytree': 0.7}. Best is trial 110 with value: 6.26.
[I 2023-08-24 10:27:57,255] Trial 111 finished with value: 6.26 and parameters: {'n_estimators': 275, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.7, 'colsample_bytree': 0.7}. Best is trial 110 with value: 6.26.
[I 2023-08-24 10:27:57,416] Trial 112 finished with value: 6.26 and parameters: {'n_estimators': 275, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.6799999999999999, 'colsample_bytree': 0.7}. Best is trial 110 with value: 6.26.
[I 2023-08-24 10:27:57,677] Trial 113 finished with value: 6.42 and parameters: {'n_estimators': 274, 'max_depth': 4, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.6799999999999999, 'colsample_bytree': 0.7}. Best is trial 110 with value: 6.26.
[I 2023-08-24 10:27:57,844] Trial 114 finished with value: 6.26 and parameters: {'n_estimators': 287, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09999999999999999, 'subsample': 0.7, 'colsample_bytree': 0.7}. Best is trial 110 with value: 6.26.
[I 2023-08-24 10:27:58,030] Trial 115 finished with value: 6.26 and parameters: {'n_estimators': 287, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.71, 'colsample_bytree': 0.7}. Best is trial 110 with value: 6.26.
[I 2023-08-24 10:27:58,208] Trial 116 finished with value: 6.26 and parameters: {'n_estimators': 287, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.71, 'colsample_bytree': 0.7}. Best is trial 110 with value: 6.26.
[I 2023-08-24 10:27:58,381] Trial 117 finished with value: 6.26 and parameters: {'n_estimators': 287, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.71, 'colsample_bytree': 0.7}. Best is trial 110 with value: 6.26.
[I 2023-08-24 10:27:58,562] Trial 118 finished with value: 6.26 and parameters: {'n_estimators': 284, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.08, 'subsample': 0.73, 'colsample_bytree': 0.7}. Best is trial 110 with value: 6.26.
[I 2023-08-24 10:27:58,740] Trial 119 finished with value: 6.3 and parameters: {'n_estimators': 286, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.06999999999999999, 'subsample': 0.71, 'colsample_bytree': 0.7}. Best is trial 110 with value: 6.26.
[I 2023-08-24 10:27:58,914] Trial 120 finished with value: 6.25 and parameters: {'n_estimators': 276, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.76, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:27:59,093] Trial 121 finished with value: 6.26 and parameters: {'n_estimators': 278, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.08, 'subsample': 0.74, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:27:59,259] Trial 122 finished with value: 6.27 and parameters: {'n_estimators': 276, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.08, 'subsample': 0.74, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:27:59,433] Trial 123 finished with value: 6.3 and parameters: {'n_estimators': 284, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.06999999999999999, 'subsample': 0.76, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:27:59,600] Trial 124 finished with value: 6.27 and parameters: {'n_estimators': 277, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.08, 'subsample': 0.73, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:27:59,777] Trial 125 finished with value: 6.27 and parameters: {'n_estimators': 280, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.08, 'subsample': 0.76, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:27:59,952] Trial 126 finished with value: 6.26 and parameters: {'n_estimators': 289, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.72, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:00,141] Trial 127 finished with value: 6.26 and parameters: {'n_estimators': 289, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.72, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:00,310] Trial 128 finished with value: 6.26 and parameters: {'n_estimators': 287, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.08, 'subsample': 0.72, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:00,484] Trial 129 finished with value: 6.26 and parameters: {'n_estimators': 288, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.08, 'subsample': 0.72, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:00,653] Trial 130 finished with value: 6.27 and parameters: {'n_estimators': 282, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.08, 'subsample': 0.72, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:00,827] Trial 131 finished with value: 6.26 and parameters: {'n_estimators': 288, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.08, 'subsample': 0.72, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:01,006] Trial 132 finished with value: 6.26 and parameters: {'n_estimators': 294, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.08, 'subsample': 0.74, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:01,178] Trial 133 finished with value: 6.3 and parameters: {'n_estimators': 288, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.06999999999999999, 'subsample': 0.71, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:01,346] Trial 134 finished with value: 6.26 and parameters: {'n_estimators': 274, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.08, 'subsample': 0.73, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:01,522] Trial 135 finished with value: 6.26 and parameters: {'n_estimators': 290, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.72, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:01,695] Trial 136 finished with value: 6.26 and parameters: {'n_estimators': 284, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.75, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:01,870] Trial 137 finished with value: 6.27 and parameters: {'n_estimators': 279, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.08, 'subsample': 0.7, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:02,050] Trial 138 finished with value: 6.26 and parameters: {'n_estimators': 295, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.73, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:02,233] Trial 139 finished with value: 6.26 and parameters: {'n_estimators': 273, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.69, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:02,404] Trial 140 finished with value: 6.26 and parameters: {'n_estimators': 285, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.08, 'subsample': 0.74, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:02,583] Trial 141 finished with value: 6.26 and parameters: {'n_estimators': 292, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.08, 'subsample': 0.72, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:02,758] Trial 142 finished with value: 6.26 and parameters: {'n_estimators': 289, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.08, 'subsample': 0.71, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:02,933] Trial 143 finished with value: 6.27 and parameters: {'n_estimators': 280, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.08, 'subsample': 0.69, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:03,120] Trial 144 finished with value: 6.26 and parameters: {'n_estimators': 289, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.7, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:03,301] Trial 145 finished with value: 6.3 and parameters: {'n_estimators': 286, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.06999999999999999, 'subsample': 0.75, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:03,474] Trial 146 finished with value: 6.26 and parameters: {'n_estimators': 298, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.72, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:03,642] Trial 147 finished with value: 6.27 and parameters: {'n_estimators': 277, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.08, 'subsample': 0.71, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:03,819] Trial 148 finished with value: 6.25 and parameters: {'n_estimators': 283, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.7, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:04,086] Trial 149 finished with value: 6.39 and parameters: {'n_estimators': 264, 'max_depth': 4, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.7, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:04,266] Trial 150 finished with value: 6.26 and parameters: {'n_estimators': 282, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.73, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:04,449] Trial 151 finished with value: 6.26 and parameters: {'n_estimators': 293, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.72, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:04,618] Trial 152 finished with value: 6.26 and parameters: {'n_estimators': 270, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.08, 'subsample': 0.7, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:04,797] Trial 153 finished with value: 6.26 and parameters: {'n_estimators': 286, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.74, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:04,973] Trial 154 finished with value: 6.25 and parameters: {'n_estimators': 276, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.71, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:05,146] Trial 155 finished with value: 6.25 and parameters: {'n_estimators': 279, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.71, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:05,316] Trial 156 finished with value: 6.25 and parameters: {'n_estimators': 275, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.69, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:05,477] Trial 157 finished with value: 6.25 and parameters: {'n_estimators': 274, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.69, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:05,651] Trial 158 finished with value: 6.25 and parameters: {'n_estimators': 274, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.69, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:05,825] Trial 159 finished with value: 6.26 and parameters: {'n_estimators': 259, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.69, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:06,002] Trial 160 finished with value: 6.26 and parameters: {'n_estimators': 272, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.69, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:06,258] Trial 161 finished with value: 6.25 and parameters: {'n_estimators': 276, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.6799999999999999, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:06,442] Trial 162 finished with value: 6.25 and parameters: {'n_estimators': 275, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.6799999999999999, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:06,614] Trial 163 finished with value: 6.26 and parameters: {'n_estimators': 268, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.6799999999999999, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:06,789] Trial 164 finished with value: 6.25 and parameters: {'n_estimators': 276, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.6799999999999999, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:06,970] Trial 165 finished with value: 6.25 and parameters: {'n_estimators': 275, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.6799999999999999, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:07,148] Trial 166 finished with value: 6.25 and parameters: {'n_estimators': 264, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.6799999999999999, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:07,317] Trial 167 finished with value: 6.25 and parameters: {'n_estimators': 264, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.6799999999999999, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:07,737] Trial 168 finished with value: 6.51 and parameters: {'n_estimators': 263, 'max_depth': 6, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.6799999999999999, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:07,910] Trial 169 finished with value: 6.26 and parameters: {'n_estimators': 266, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.67, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:08,087] Trial 170 finished with value: 6.26 and parameters: {'n_estimators': 273, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.6799999999999999, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:08,273] Trial 171 finished with value: 6.25 and parameters: {'n_estimators': 276, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.69, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:08,460] Trial 172 finished with value: 6.26 and parameters: {'n_estimators': 270, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.69, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:08,633] Trial 173 finished with value: 6.25 and parameters: {'n_estimators': 275, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.67, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:08,807] Trial 174 finished with value: 6.25 and parameters: {'n_estimators': 275, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.67, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:08,983] Trial 175 finished with value: 6.26 and parameters: {'n_estimators': 260, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.67, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:09,160] Trial 176 finished with value: 6.26 and parameters: {'n_estimators': 267, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.69, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:09,340] Trial 177 finished with value: 6.25 and parameters: {'n_estimators': 274, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.67, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:09,504] Trial 178 finished with value: 6.26 and parameters: {'n_estimators': 270, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.67, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:09,674] Trial 179 finished with value: 6.25 and parameters: {'n_estimators': 277, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.6799999999999999, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:09,853] Trial 180 finished with value: 6.25 and parameters: {'n_estimators': 279, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.67, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:10,027] Trial 181 finished with value: 6.25 and parameters: {'n_estimators': 278, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.67, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:10,198] Trial 182 finished with value: 6.25 and parameters: {'n_estimators': 264, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.6799999999999999, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:10,371] Trial 183 finished with value: 6.26 and parameters: {'n_estimators': 273, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.67, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:10,557] Trial 184 finished with value: 6.25 and parameters: {'n_estimators': 280, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.66, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:10,733] Trial 185 finished with value: 6.25 and parameters: {'n_estimators': 276, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.6799999999999999, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:10,912] Trial 186 finished with value: 6.26 and parameters: {'n_estimators': 268, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.6799999999999999, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:11,088] Trial 187 finished with value: 6.26 and parameters: {'n_estimators': 257, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.69, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:11,273] Trial 188 finished with value: 6.25 and parameters: {'n_estimators': 281, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.69, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:11,610] Trial 189 finished with value: 6.51 and parameters: {'n_estimators': 272, 'max_depth': 5, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.66, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:11,787] Trial 190 finished with value: 6.25 and parameters: {'n_estimators': 276, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.67, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:11,971] Trial 191 finished with value: 6.25 and parameters: {'n_estimators': 278, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.67, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:12,151] Trial 192 finished with value: 6.25 and parameters: {'n_estimators': 279, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.6799999999999999, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:12,327] Trial 193 finished with value: 6.26 and parameters: {'n_estimators': 271, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.66, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:12,504] Trial 194 finished with value: 6.26 and parameters: {'n_estimators': 267, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.6799999999999999, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:12,697] Trial 195 finished with value: 6.25 and parameters: {'n_estimators': 276, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.67, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:12,879] Trial 196 finished with value: 6.25 and parameters: {'n_estimators': 281, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.69, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:13,051] Trial 197 finished with value: 6.25 and parameters: {'n_estimators': 264, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.67, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:13,232] Trial 198 finished with value: 6.26 and parameters: {'n_estimators': 273, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.69, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.
[I 2023-08-24 10:28:13,406] Trial 199 finished with value: 6.25 and parameters: {'n_estimators': 278, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.67, 'colsample_bytree': 0.7}. Best is trial 120 with value: 6.25.

Лучшие параметры: {'n_estimators': 276, 'max_depth': 2, 'min_child_samples': 7, 'learning_rate': 0.09, 'subsample': 0.76, 'colsample_bytree': 0.7}
Лучшее значение метрики: 6.25

<timed exec>:57: ExperimentalWarning:

plot_param_importances is experimental (supported from v2.2.0). The interface can change in the future.

<timed exec>:73: ExperimentalWarning:

plot_optimization_history is experimental (supported from v2.2.0). The interface can change in the future.

Wall time: 43.8 s


# Создаём график контура
fig = optuna.visualization.plot_contour(study, params=["max_depth", "n_estimators"])
fig.show()


lgbm

LGBMRegressor(colsample_bytree=0.7, learning_rate=0.09, max_depth=2,
              min_child_samples=7, n_estimators=276, random_state=140823,
              subsample=0.76)


%%time

def objective(trial):
    # Определяем параметры для подбора
    iterations = trial.suggest_int('iterations', 100, 500)
    depth = trial.suggest_int('depth', 1, 10)
    learning_rate = trial.suggest_float('learning_rate', 0.01, 0.1, step=0.01)

# Создаем модель с определенными параметрами
    model = CatBoostRegressor(iterations=iterations,
                              depth=depth,
                              learning_rate=learning_rate,
                              random_state=RANDOM_STATE,
                              silent=True)

# Оцениваем качество модели с помощью кросс-валидации
    scores = round(abs(cross_val_score(model,
                                      features_train,
                                      target_train,
                                      cv=kf,
                                      scoring='neg_mean_absolute_error', n_jobs=-1)).mean(), 2)

# Возвращаем среднее значение метрики
    return abs(scores.mean())

# Создаем объект Study и запускаем оптимизацию
study = optuna.create_study(direction='minimize')
study.optimize(objective, n_trials=50)

print('Лучшие параметры:', study.best_params)
print('Лучшее значение метрики:', study.best_value)

# Создаем датафрейм с значениями гиперпараметров и метрикой
results = pd.DataFrame({'iterations': [],
                       'depth': [],
                        'learning_rate': [],
                        'score': []})
# Заполняем данными из best_trials
for trial in study.best_trials:
    params = trial.params
    params['score'] = trial.value
    results = results.append(params, ignore_index=True)

# Сортируем датафрейм по убыванию значения метрики
results = results.sort_values(by='score', ascending=False)

# Выводим 10 лучших моделей
display(results.head(10))

# Создаем график важности параметров
optuna.visualization.matplotlib.plot_param_importances(study)
plt.xlabel('Важность гиперпараметров',
           fontsize=12,
           color='DarkSlateGray')
plt.ylabel('Гиперпараметры', fontsize=12, color='DarkSlateGray')
plt.suptitle('Важность гиперпараметров', fontsize=15, color='DarkSlateGray')
plt.minorticks_on()
plt.grid(which='minor', linestyle=':')
plt.grid(True)

# Создаем график истории оптимизации
optuna.visualization.matplotlib.plot_optimization_history(study)
plt.tight_layout()
plt.xlabel('Итерации', fontsize=12, color='DarkSlateGray')
plt.ylabel('Целевая метрика', fontsize=12, color='DarkSlateGray')
plt.suptitle('История оптимизации \n \n', fontsize=15, color='DarkSlateGray')
plt.minorticks_on()
plt.grid(which='minor', linestyle=':')
plt.grid(True)
plt.show()

[I 2023-08-24 10:28:18,439] A new study created in memory with name: no-name-040408ae-6f0a-42fc-88b3-1dc1a49dc00d
[I 2023-08-24 10:28:21,548] Trial 0 finished with value: 6.37 and parameters: {'iterations': 480, 'depth': 4, 'learning_rate': 0.09}. Best is trial 0 with value: 6.37.
[I 2023-08-24 10:28:25,071] Trial 1 finished with value: 6.31 and parameters: {'iterations': 205, 'depth': 7, 'learning_rate': 0.06999999999999999}. Best is trial 1 with value: 6.31.
[I 2023-08-24 10:28:25,277] Trial 2 finished with value: 6.67 and parameters: {'iterations': 114, 'depth': 2, 'learning_rate': 0.060000000000000005}. Best is trial 1 with value: 6.31.
[I 2023-08-24 10:28:25,708] Trial 3 finished with value: 6.55 and parameters: {'iterations': 423, 'depth': 1, 'learning_rate': 0.060000000000000005}. Best is trial 1 with value: 6.31.
[I 2023-08-24 10:28:27,567] Trial 4 finished with value: 6.33 and parameters: {'iterations': 206, 'depth': 6, 'learning_rate': 0.06999999999999999}. Best is trial 1 with value: 6.31.
[I 2023-08-24 10:29:30,281] Trial 5 finished with value: 6.42 and parameters: {'iterations': 464, 'depth': 10, 'learning_rate': 0.03}. Best is trial 1 with value: 6.31.
[I 2023-08-24 10:29:31,125] Trial 6 finished with value: 6.45 and parameters: {'iterations': 349, 'depth': 3, 'learning_rate': 0.03}. Best is trial 1 with value: 6.31.
[I 2023-08-24 10:29:34,805] Trial 7 finished with value: 6.46 and parameters: {'iterations': 416, 'depth': 6, 'learning_rate': 0.09}. Best is trial 1 with value: 6.31.
[I 2023-08-24 10:29:37,056] Trial 8 finished with value: 6.41 and parameters: {'iterations': 400, 'depth': 5, 'learning_rate': 0.09999999999999999}. Best is trial 1 with value: 6.31.
[I 2023-08-24 10:29:37,737] Trial 9 finished with value: 6.93 and parameters: {'iterations': 181, 'depth': 4, 'learning_rate': 0.01}. Best is trial 1 with value: 6.31.
[I 2023-08-24 10:29:52,728] Trial 10 finished with value: 6.37 and parameters: {'iterations': 269, 'depth': 9, 'learning_rate': 0.04}. Best is trial 1 with value: 6.31.
[I 2023-08-24 10:29:56,267] Trial 11 finished with value: 6.32 and parameters: {'iterations': 233, 'depth': 7, 'learning_rate': 0.06999999999999999}. Best is trial 1 with value: 6.31.
[I 2023-08-24 10:30:03,344] Trial 12 finished with value: 6.41 and parameters: {'iterations': 258, 'depth': 8, 'learning_rate': 0.08}. Best is trial 1 with value: 6.31.
[I 2023-08-24 10:30:06,201] Trial 13 finished with value: 6.33 and parameters: {'iterations': 186, 'depth': 7, 'learning_rate': 0.05}. Best is trial 1 with value: 6.31.
[I 2023-08-24 10:30:09,510] Trial 14 finished with value: 6.38 and parameters: {'iterations': 111, 'depth': 8, 'learning_rate': 0.06999999999999999}. Best is trial 1 with value: 6.31.
[I 2023-08-24 10:30:14,343] Trial 15 finished with value: 6.33 and parameters: {'iterations': 316, 'depth': 7, 'learning_rate': 0.08}. Best is trial 1 with value: 6.31.
[I 2023-08-24 10:30:45,289] Trial 16 finished with value: 6.42 and parameters: {'iterations': 231, 'depth': 10, 'learning_rate': 0.05}. Best is trial 1 with value: 6.31.
[I 2023-08-24 10:30:49,579] Trial 17 finished with value: 6.36 and parameters: {'iterations': 152, 'depth': 8, 'learning_rate': 0.06999999999999999}. Best is trial 1 with value: 6.31.
[I 2023-08-24 10:30:51,302] Trial 18 finished with value: 6.36 and parameters: {'iterations': 307, 'depth': 5, 'learning_rate': 0.09999999999999999}. Best is trial 1 with value: 6.31.
[I 2023-08-24 10:30:55,087] Trial 19 finished with value: 6.31 and parameters: {'iterations': 247, 'depth': 7, 'learning_rate': 0.08}. Best is trial 1 with value: 6.31.
[I 2023-08-24 10:31:13,711] Trial 20 finished with value: 6.43 and parameters: {'iterations': 340, 'depth': 9, 'learning_rate': 0.09}. Best is trial 1 with value: 6.31.
[I 2023-08-24 10:31:17,624] Trial 21 finished with value: 6.31 and parameters: {'iterations': 255, 'depth': 7, 'learning_rate': 0.08}. Best is trial 1 with value: 6.31.
[I 2023-08-24 10:31:20,138] Trial 22 finished with value: 6.38 and parameters: {'iterations': 272, 'depth': 6, 'learning_rate': 0.08}. Best is trial 1 with value: 6.31.
[I 2023-08-24 10:31:27,129] Trial 23 finished with value: 6.41 and parameters: {'iterations': 227, 'depth': 8, 'learning_rate': 0.08}. Best is trial 1 with value: 6.31.
[I 2023-08-24 10:31:29,710] Trial 24 finished with value: 6.32 and parameters: {'iterations': 154, 'depth': 7, 'learning_rate': 0.060000000000000005}. Best is trial 1 with value: 6.31.
[I 2023-08-24 10:31:45,244] Trial 25 finished with value: 6.42 and parameters: {'iterations': 281, 'depth': 9, 'learning_rate': 0.09}. Best is trial 1 with value: 6.31.
[I 2023-08-24 10:31:46,175] Trial 26 finished with value: 6.31 and parameters: {'iterations': 158, 'depth': 5, 'learning_rate': 0.09999999999999999}. Best is trial 1 with value: 6.31.
[I 2023-08-24 10:31:49,293] Trial 27 finished with value: 6.36 and parameters: {'iterations': 352, 'depth': 6, 'learning_rate': 0.06999999999999999}. Best is trial 1 with value: 6.31.
[I 2023-08-24 10:31:53,067] Trial 28 finished with value: 6.3 and parameters: {'iterations': 239, 'depth': 7, 'learning_rate': 0.08}. Best is trial 28 with value: 6.3.
[I 2023-08-24 10:31:53,795] Trial 29 finished with value: 6.42 and parameters: {'iterations': 189, 'depth': 4, 'learning_rate': 0.05}. Best is trial 28 with value: 6.3.
[I 2023-08-24 10:32:05,257] Trial 30 finished with value: 6.4 and parameters: {'iterations': 211, 'depth': 9, 'learning_rate': 0.09}. Best is trial 28 with value: 6.3.
[I 2023-08-24 10:32:09,071] Trial 31 finished with value: 6.31 and parameters: {'iterations': 246, 'depth': 7, 'learning_rate': 0.08}. Best is trial 28 with value: 6.3.
[I 2023-08-24 10:32:17,155] Trial 32 finished with value: 6.44 and parameters: {'iterations': 284, 'depth': 8, 'learning_rate': 0.08}. Best is trial 28 with value: 6.3.
[I 2023-08-24 10:32:20,555] Trial 33 finished with value: 6.3 and parameters: {'iterations': 212, 'depth': 7, 'learning_rate': 0.060000000000000005}. Best is trial 28 with value: 6.3.
[I 2023-08-24 10:32:21,771] Trial 34 finished with value: 6.39 and parameters: {'iterations': 133, 'depth': 6, 'learning_rate': 0.060000000000000005}. Best is trial 28 with value: 6.3.
[I 2023-08-24 10:32:22,048] Trial 35 finished with value: 6.66 and parameters: {'iterations': 209, 'depth': 1, 'learning_rate': 0.060000000000000005}. Best is trial 28 with value: 6.3.
[I 2023-08-24 10:32:23,682] Trial 36 finished with value: 6.32 and parameters: {'iterations': 180, 'depth': 6, 'learning_rate': 0.06999999999999999}. Best is trial 28 with value: 6.3.
[I 2023-08-24 10:32:24,885] Trial 37 finished with value: 6.39 and parameters: {'iterations': 210, 'depth': 5, 'learning_rate': 0.04}. Best is trial 28 with value: 6.3.
[I 2023-08-24 10:32:25,682] Trial 38 finished with value: 6.37 and parameters: {'iterations': 325, 'depth': 3, 'learning_rate': 0.06999999999999999}. Best is trial 28 with value: 6.3.
[I 2023-08-24 10:32:31,469] Trial 39 finished with value: 6.33 and parameters: {'iterations': 371, 'depth': 7, 'learning_rate': 0.060000000000000005}. Best is trial 28 with value: 6.3.
[I 2023-08-24 10:32:44,494] Trial 40 finished with value: 6.34 and parameters: {'iterations': 465, 'depth': 8, 'learning_rate': 0.02}. Best is trial 28 with value: 6.3.
[I 2023-08-24 10:32:48,266] Trial 41 finished with value: 6.31 and parameters: {'iterations': 246, 'depth': 7, 'learning_rate': 0.08}. Best is trial 28 with value: 6.3.
[I 2023-08-24 10:32:50,928] Trial 42 finished with value: 6.4 and parameters: {'iterations': 296, 'depth': 6, 'learning_rate': 0.09}. Best is trial 28 with value: 6.3.
[I 2023-08-24 10:32:54,645] Trial 43 finished with value: 6.31 and parameters: {'iterations': 242, 'depth': 7, 'learning_rate': 0.060000000000000005}. Best is trial 28 with value: 6.3.
[I 2023-08-24 10:33:01,737] Trial 44 finished with value: 6.39 and parameters: {'iterations': 260, 'depth': 8, 'learning_rate': 0.06999999999999999}. Best is trial 28 with value: 6.3.
[I 2023-08-24 10:33:05,119] Trial 45 finished with value: 6.3 and parameters: {'iterations': 221, 'depth': 7, 'learning_rate': 0.08}. Best is trial 28 with value: 6.3.
[I 2023-08-24 10:33:06,951] Trial 46 finished with value: 6.32 and parameters: {'iterations': 198, 'depth': 6, 'learning_rate': 0.06999999999999999}. Best is trial 28 with value: 6.3.
[I 2023-08-24 10:33:11,526] Trial 47 finished with value: 6.41 and parameters: {'iterations': 163, 'depth': 8, 'learning_rate': 0.04}. Best is trial 28 with value: 6.3.
[I 2023-08-24 10:33:23,670] Trial 48 finished with value: 6.4 and parameters: {'iterations': 219, 'depth': 9, 'learning_rate': 0.09}. Best is trial 28 with value: 6.3.
[I 2023-08-24 10:33:24,694] Trial 49 finished with value: 6.38 and parameters: {'iterations': 174, 'depth': 5, 'learning_rate': 0.05}. Best is trial 28 with value: 6.3.

Лучшие параметры: {'iterations': 239, 'depth': 7, 'learning_rate': 0.08}
Лучшее значение метрики: 6.3

<timed exec>:49: ExperimentalWarning:

plot_param_importances is experimental (supported from v2.2.0). The interface can change in the future.

<timed exec>:60: ExperimentalWarning:

plot_optimization_history is experimental (supported from v2.2.0). The interface can change in the future.

Wall time: 5min 8s


# Создаём график контура
fig = optuna.visualization.plot_contour(study, params=["iterations", "depth"])
fig.show()


# Создаём график контура
fig = optuna.visualization.plot_contour(study, params=["iterations", "learning_rate"])
fig.show()


%%time

def objective(trial):
    # Определяем параметры для подбора
    iterations = trial.suggest_int('iterations', 250, 600)
    depth = trial.suggest_int('depth', 2, 8)
    learning_rate = trial.suggest_float('learning_rate', 0.01, 0.08, step=0.01)
# Создаем модель с определенными параметрами
    model = CatBoostRegressor(iterations=iterations,
                              depth=depth,
                              learning_rate=learning_rate,
                              random_state=RANDOM_STATE,
                              silent=True)

# Оцениваем качество модели с помощью кросс-валидации
    scores = round(abs(cross_val_score(model,
                                      features_train,
                                      target_train,
                                      cv=kf,
                                      scoring='neg_mean_absolute_error', n_jobs=-1)).mean(), 2)

# Возвращаем среднее значение метрики
    return abs(scores.mean())

# Создаем объект Study и запускаем оптимизацию
study = optuna.create_study(direction='minimize')
study.optimize(objective, n_trials=50)

print('Лучшие параметры:', study.best_params)
print('Лучшее значение метрики:', study.best_value)

# Создаем датафрейм с значениями гиперпараметров и метрикой
results = pd.DataFrame({'iterations': [],
                       'depth': [],
                        'learning_rate': [],
                        'score': []})
# Заполняем данными из best_trials
for trial in study.best_trials:
    params = trial.params
    params['score'] = trial.value
    results = results.append(params, ignore_index=True)

# Сортируем датафрейм по убыванию значения метрики
results = results.sort_values(by='score', ascending=False)

# Выводим 10 лучших моделей
display(results.head(10))

# Создаем график важности параметров
optuna.visualization.matplotlib.plot_param_importances(study)
plt.xlabel('Важность гиперпараметров',
           fontsize=12,
           color='DarkSlateGray')
plt.ylabel('Гиперпараметры',
           fontsize=12,
           color='DarkSlateGray')
plt.suptitle('Важность гиперпараметров',
             fontsize=15,
             color='DarkSlateGray')
plt.minorticks_on()
plt.grid(which='minor',
         linestyle=':')
plt.grid(True)

# Создаем график истории оптимизации
optuna.visualization.matplotlib.plot_optimization_history(study)
plt.tight_layout()
plt.xlabel('Итерации',
           fontsize=12,
           color='DarkSlateGray')
plt.ylabel('Целевая метрика',
           fontsize=12,
           color='DarkSlateGray')
plt.suptitle('История оптимизации \n \n',
             fontsize=15,
             color='DarkSlateGray')
plt.minorticks_on()
plt.grid(which='minor',
         linestyle=':')
plt.grid(True)
plt.show()

[I 2023-08-24 10:33:26,960] A new study created in memory with name: no-name-b3be6fe5-b320-4c8d-af84-dc7eed094f89
[I 2023-08-24 10:33:31,650] Trial 0 finished with value: 6.3 and parameters: {'iterations': 500, 'depth': 6, 'learning_rate': 0.04}. Best is trial 0 with value: 6.3.
[I 2023-08-24 10:33:32,390] Trial 1 finished with value: 6.36 and parameters: {'iterations': 469, 'depth': 2, 'learning_rate': 0.060000000000000005}. Best is trial 0 with value: 6.3.
[I 2023-08-24 10:33:35,262] Trial 2 finished with value: 6.53 and parameters: {'iterations': 517, 'depth': 5, 'learning_rate': 0.01}. Best is trial 0 with value: 6.3.
[I 2023-08-24 10:33:37,968] Trial 3 finished with value: 6.34 and parameters: {'iterations': 306, 'depth': 6, 'learning_rate': 0.06999999999999999}. Best is trial 0 with value: 6.3.
[I 2023-08-24 10:33:45,167] Trial 4 finished with value: 6.31 and parameters: {'iterations': 475, 'depth': 7, 'learning_rate': 0.03}. Best is trial 0 with value: 6.3.
[I 2023-08-24 10:33:48,370] Trial 5 finished with value: 6.41 and parameters: {'iterations': 595, 'depth': 5, 'learning_rate': 0.08}. Best is trial 0 with value: 6.3.
[I 2023-08-24 10:33:49,265] Trial 6 finished with value: 6.42 and parameters: {'iterations': 596, 'depth': 2, 'learning_rate': 0.03}. Best is trial 0 with value: 6.3.
[I 2023-08-24 10:33:57,470] Trial 7 finished with value: 6.35 and parameters: {'iterations': 488, 'depth': 7, 'learning_rate': 0.060000000000000005}. Best is trial 0 with value: 6.3.
[I 2023-08-24 10:33:59,474] Trial 8 finished with value: 6.3 and parameters: {'iterations': 353, 'depth': 5, 'learning_rate': 0.06999999999999999}. Best is trial 0 with value: 6.3.
[I 2023-08-24 10:34:00,392] Trial 9 finished with value: 6.36 and parameters: {'iterations': 570, 'depth': 2, 'learning_rate': 0.060000000000000005}. Best is trial 0 with value: 6.3.
[I 2023-08-24 10:34:10,864] Trial 10 finished with value: 6.32 and parameters: {'iterations': 392, 'depth': 8, 'learning_rate': 0.04}. Best is trial 0 with value: 6.3.
[I 2023-08-24 10:34:12,171] Trial 11 finished with value: 6.32 and parameters: {'iterations': 364, 'depth': 4, 'learning_rate': 0.04}. Best is trial 0 with value: 6.3.
[I 2023-08-24 10:34:13,256] Trial 12 finished with value: 6.3 and parameters: {'iterations': 290, 'depth': 4, 'learning_rate': 0.08}. Best is trial 0 with value: 6.3.
[I 2023-08-24 10:34:16,508] Trial 13 finished with value: 6.59 and parameters: {'iterations': 367, 'depth': 6, 'learning_rate': 0.01}. Best is trial 0 with value: 6.3.
[I 2023-08-24 10:34:17,567] Trial 14 finished with value: 6.34 and parameters: {'iterations': 254, 'depth': 4, 'learning_rate': 0.05}. Best is trial 0 with value: 6.3.
[I 2023-08-24 10:34:21,229] Trial 15 finished with value: 6.3 and parameters: {'iterations': 407, 'depth': 6, 'learning_rate': 0.03}. Best is trial 0 with value: 6.3.
[I 2023-08-24 10:34:22,288] Trial 16 finished with value: 6.35 and parameters: {'iterations': 440, 'depth': 3, 'learning_rate': 0.05}. Best is trial 0 with value: 6.3.
[I 2023-08-24 10:34:37,140] Trial 17 finished with value: 6.47 and parameters: {'iterations': 533, 'depth': 8, 'learning_rate': 0.06999999999999999}. Best is trial 0 with value: 6.3.
[I 2023-08-24 10:34:42,324] Trial 18 finished with value: 6.39 and parameters: {'iterations': 340, 'depth': 7, 'learning_rate': 0.02}. Best is trial 0 with value: 6.3.
[I 2023-08-24 10:34:44,660] Trial 19 finished with value: 6.32 and parameters: {'iterations': 437, 'depth': 5, 'learning_rate': 0.06999999999999999}. Best is trial 0 with value: 6.3.
[I 2023-08-24 10:34:49,066] Trial 20 finished with value: 6.35 and parameters: {'iterations': 530, 'depth': 6, 'learning_rate': 0.05}. Best is trial 0 with value: 6.3.
[I 2023-08-24 10:34:50,113] Trial 21 finished with value: 6.29 and parameters: {'iterations': 278, 'depth': 4, 'learning_rate': 0.08}. Best is trial 21 with value: 6.29.
[I 2023-08-24 10:34:50,875] Trial 22 finished with value: 6.36 and parameters: {'iterations': 305, 'depth': 3, 'learning_rate': 0.08}. Best is trial 21 with value: 6.29.
[I 2023-08-24 10:34:51,830] Trial 23 finished with value: 6.3 and parameters: {'iterations': 252, 'depth': 4, 'learning_rate': 0.06999999999999999}. Best is trial 21 with value: 6.29.
[I 2023-08-24 10:34:52,620] Trial 24 finished with value: 6.37 and parameters: {'iterations': 325, 'depth': 3, 'learning_rate': 0.08}. Best is trial 21 with value: 6.29.
[I 2023-08-24 10:34:54,179] Trial 25 finished with value: 6.32 and parameters: {'iterations': 279, 'depth': 5, 'learning_rate': 0.04}. Best is trial 21 with value: 6.29.
[I 2023-08-24 10:34:56,086] Trial 26 finished with value: 6.29 and parameters: {'iterations': 346, 'depth': 5, 'learning_rate': 0.060000000000000005}. Best is trial 21 with value: 6.29.
[I 2023-08-24 10:34:59,662] Trial 27 finished with value: 6.33 and parameters: {'iterations': 392, 'depth': 6, 'learning_rate': 0.060000000000000005}. Best is trial 21 with value: 6.29.
[I 2023-08-24 10:35:00,668] Trial 28 finished with value: 6.34 and parameters: {'iterations': 275, 'depth': 4, 'learning_rate': 0.05}. Best is trial 21 with value: 6.29.
[I 2023-08-24 10:35:05,587] Trial 29 finished with value: 6.33 and parameters: {'iterations': 320, 'depth': 7, 'learning_rate': 0.060000000000000005}. Best is trial 21 with value: 6.29.
[I 2023-08-24 10:35:06,812] Trial 30 finished with value: 6.48 and parameters: {'iterations': 452, 'depth': 3, 'learning_rate': 0.02}. Best is trial 21 with value: 6.29.
[I 2023-08-24 10:35:08,757] Trial 31 finished with value: 6.3 and parameters: {'iterations': 351, 'depth': 5, 'learning_rate': 0.06999999999999999}. Best is trial 21 with value: 6.29.
[I 2023-08-24 10:35:11,009] Trial 32 finished with value: 6.32 and parameters: {'iterations': 410, 'depth': 5, 'learning_rate': 0.06999999999999999}. Best is trial 21 with value: 6.29.
[I 2023-08-24 10:35:13,064] Trial 33 finished with value: 6.34 and parameters: {'iterations': 379, 'depth': 5, 'learning_rate': 0.08}. Best is trial 21 with value: 6.29.
[I 2023-08-24 10:35:17,532] Trial 34 finished with value: 6.35 and parameters: {'iterations': 503, 'depth': 6, 'learning_rate': 0.060000000000000005}. Best is trial 21 with value: 6.29.
[I 2023-08-24 10:35:18,730] Trial 35 finished with value: 6.32 and parameters: {'iterations': 336, 'depth': 4, 'learning_rate': 0.06999999999999999}. Best is trial 21 with value: 6.29.
[I 2023-08-24 10:35:20,414] Trial 36 finished with value: 6.32 and parameters: {'iterations': 304, 'depth': 5, 'learning_rate': 0.08}. Best is trial 21 with value: 6.29.
[I 2023-08-24 10:35:24,741] Trial 37 finished with value: 6.35 and parameters: {'iterations': 478, 'depth': 6, 'learning_rate': 0.060000000000000005}. Best is trial 21 with value: 6.29.
[I 2023-08-24 10:35:26,734] Trial 38 finished with value: 6.32 and parameters: {'iterations': 560, 'depth': 4, 'learning_rate': 0.04}. Best is trial 21 with value: 6.29.
[I 2023-08-24 10:35:29,274] Trial 39 finished with value: 6.32 and parameters: {'iterations': 459, 'depth': 5, 'learning_rate': 0.06999999999999999}. Best is trial 21 with value: 6.29.
[I 2023-08-24 10:35:35,687] Trial 40 finished with value: 6.3 and parameters: {'iterations': 422, 'depth': 7, 'learning_rate': 0.03}. Best is trial 21 with value: 6.29.
[I 2023-08-24 10:35:36,710] Trial 41 finished with value: 6.3 and parameters: {'iterations': 286, 'depth': 4, 'learning_rate': 0.08}. Best is trial 21 with value: 6.29.
[I 2023-08-24 10:35:37,792] Trial 42 finished with value: 6.3 and parameters: {'iterations': 294, 'depth': 4, 'learning_rate': 0.08}. Best is trial 21 with value: 6.29.
[I 2023-08-24 10:35:39,573] Trial 43 finished with value: 6.32 and parameters: {'iterations': 318, 'depth': 5, 'learning_rate': 0.08}. Best is trial 21 with value: 6.29.
[I 2023-08-24 10:35:40,260] Trial 44 finished with value: 6.39 and parameters: {'iterations': 268, 'depth': 3, 'learning_rate': 0.06999999999999999}. Best is trial 21 with value: 6.29.
[I 2023-08-24 10:35:41,542] Trial 45 finished with value: 6.32 and parameters: {'iterations': 351, 'depth': 4, 'learning_rate': 0.08}. Best is trial 21 with value: 6.29.
[I 2023-08-24 10:35:43,948] Trial 46 finished with value: 6.32 and parameters: {'iterations': 268, 'depth': 6, 'learning_rate': 0.060000000000000005}. Best is trial 21 with value: 6.29.
[I 2023-08-24 10:35:44,546] Trial 47 finished with value: 6.38 and parameters: {'iterations': 367, 'depth': 2, 'learning_rate': 0.05}. Best is trial 21 with value: 6.29.
[I 2023-08-24 10:35:46,237] Trial 48 finished with value: 6.29 and parameters: {'iterations': 305, 'depth': 5, 'learning_rate': 0.06999999999999999}. Best is trial 21 with value: 6.29.
[I 2023-08-24 10:35:49,188] Trial 49 finished with value: 6.29 and parameters: {'iterations': 329, 'depth': 6, 'learning_rate': 0.04}. Best is trial 21 with value: 6.29.

Лучшие параметры: {'iterations': 278, 'depth': 4, 'learning_rate': 0.08}
Лучшее значение метрики: 6.29

<timed exec>:48: ExperimentalWarning:

plot_param_importances is experimental (supported from v2.2.0). The interface can change in the future.

<timed exec>:64: ExperimentalWarning:

plot_optimization_history is experimental (supported from v2.2.0). The interface can change in the future.

Wall time: 2min 24s


# Создаём график контура
fig = optuna.visualization.plot_contour(study, params=["iterations", "depth"])
fig.show()


# Создаём график контура
fig = optuna.visualization.plot_contour(study, params=["iterations", "learning_rate"])
fig.show()


%%time

def objective(trial):
    # Определяем параметры для подбора
    iterations = trial.suggest_int('iterations', 300, 500)
    depth = trial.suggest_int('depth', 5, 7)
    learning_rate = trial.suggest_float('learning_rate', 0.02, 0.06, step=0.01)

# Создаем модель с определенными параметрами
    model = CatBoostRegressor(iterations=iterations,
                              depth=depth,
                              learning_rate=learning_rate,
                              random_state=RANDOM_STATE,
                              silent=True)

# Оцениваем качество модели с помощью кросс-валидации
    scores = round(abs(cross_val_score(model,
                                      features_train,
                                      target_train,
                                      cv=kf,
                                      scoring='neg_mean_absolute_error', n_jobs=-1)).mean(), 2)

# Возвращаем среднее значение метрики
    return abs(scores.mean())

# Создаем объект Study и запускаем оптимизацию
study = optuna.create_study(direction='minimize')
study.optimize(objective, n_trials=200)

print('Лучшие параметры:', study.best_params)
print('Лучшее значение метрики:', study.best_value)

# Создаем датафрейм с значениями гиперпараметров и метрикой
results = pd.DataFrame({'iterations': [],
                       'depth': [],
                        'learning_rate': [],
                        'score': []})
# Заполняем данными из best_trials
for trial in study.best_trials:
    params = trial.params
    params['score'] = trial.value
    results = results.append(params, ignore_index=True)

# Сортируем датафрейм по убыванию значения метрики
results = results.sort_values(by='score', ascending=False)

# Выводим 10 лучших моделей
display(results.head(10))

# Создаем график важности параметров
optuna.visualization.matplotlib.plot_param_importances(study)
plt.xlabel('Важность гиперпараметров',
           fontsize=12,
           color='DarkSlateGray')
plt.ylabel('Гиперпараметры',
           fontsize=12,
           color='DarkSlateGray')
plt.suptitle('Важность гиперпараметров',
             fontsize=15,
             color='DarkSlateGray')
plt.minorticks_on()
plt.grid(which='minor',
         linestyle=':')
plt.grid(True)

# Создаем график истории оптимизации
optuna.visualization.matplotlib.plot_optimization_history(study)
plt.tight_layout()
plt.xlabel('Итерации',
           fontsize=12,
           color='DarkSlateGray')
plt.ylabel('Целевая метрика',
           fontsize=12,
           color='DarkSlateGray')
plt.suptitle('История оптимизации \n \n',
             fontsize=15,
             color='DarkSlateGray')
plt.minorticks_on()
plt.grid(which='minor',
         linestyle=':')
plt.grid(True)
plt.show()

# Сохранение модели с наилучшими параметрами
catboost = CatBoostRegressor(iterations=study.best_params['iterations'],
                             depth=study.best_params['depth'],
                             learning_rate=study.best_params['learning_rate'],
                             random_state=RANDOM_STATE,
                             silent=True)

[I 2023-08-24 10:35:51,483] A new study created in memory with name: no-name-b1ca16d1-9e93-4d5a-bbe9-39ec2bed48d2
[I 2023-08-24 10:35:54,756] Trial 0 finished with value: 6.33 and parameters: {'iterations': 333, 'depth': 6, 'learning_rate': 0.06}. Best is trial 0 with value: 6.33.
[I 2023-08-24 10:35:58,415] Trial 1 finished with value: 6.35 and parameters: {'iterations': 416, 'depth': 6, 'learning_rate': 0.02}. Best is trial 0 with value: 6.33.
[I 2023-08-24 10:36:00,402] Trial 2 finished with value: 6.32 and parameters: {'iterations': 361, 'depth': 5, 'learning_rate': 0.03}. Best is trial 2 with value: 6.32.
[I 2023-08-24 10:36:04,443] Trial 3 finished with value: 6.29 and parameters: {'iterations': 461, 'depth': 6, 'learning_rate': 0.04}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:36:11,497] Trial 4 finished with value: 6.3 and parameters: {'iterations': 484, 'depth': 7, 'learning_rate': 0.03}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:36:13,929] Trial 5 finished with value: 6.3 and parameters: {'iterations': 446, 'depth': 5, 'learning_rate': 0.04}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:36:15,953] Trial 6 finished with value: 6.3 and parameters: {'iterations': 374, 'depth': 5, 'learning_rate': 0.04}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:36:19,679] Trial 7 finished with value: 6.34 and parameters: {'iterations': 422, 'depth': 6, 'learning_rate': 0.02}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:36:26,161] Trial 8 finished with value: 6.36 and parameters: {'iterations': 383, 'depth': 7, 'learning_rate': 0.02}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:36:28,305] Trial 9 finished with value: 6.33 and parameters: {'iterations': 344, 'depth': 5, 'learning_rate': 0.03}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:36:35,955] Trial 10 finished with value: 6.35 and parameters: {'iterations': 497, 'depth': 7, 'learning_rate': 0.06}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:36:44,235] Trial 11 finished with value: 6.33 and parameters: {'iterations': 491, 'depth': 7, 'learning_rate': 0.05}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:36:52,194] Trial 12 finished with value: 6.31 and parameters: {'iterations': 467, 'depth': 7, 'learning_rate': 0.03}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:36:56,507] Trial 13 finished with value: 6.34 and parameters: {'iterations': 457, 'depth': 6, 'learning_rate': 0.05}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:37:00,700] Trial 14 finished with value: 6.29 and parameters: {'iterations': 437, 'depth': 6, 'learning_rate': 0.04}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:37:04,969] Trial 15 finished with value: 6.34 and parameters: {'iterations': 436, 'depth': 6, 'learning_rate': 0.05}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:37:08,912] Trial 16 finished with value: 6.29 and parameters: {'iterations': 407, 'depth': 6, 'learning_rate': 0.04}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:37:11,718] Trial 17 finished with value: 6.33 and parameters: {'iterations': 301, 'depth': 6, 'learning_rate': 0.05}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:37:14,232] Trial 18 finished with value: 6.3 and parameters: {'iterations': 462, 'depth': 5, 'learning_rate': 0.04}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:37:18,474] Trial 19 finished with value: 6.29 and parameters: {'iterations': 434, 'depth': 6, 'learning_rate': 0.04}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:37:22,409] Trial 20 finished with value: 6.34 and parameters: {'iterations': 395, 'depth': 6, 'learning_rate': 0.05}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:37:26,147] Trial 21 finished with value: 6.29 and parameters: {'iterations': 413, 'depth': 6, 'learning_rate': 0.04}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:37:29,724] Trial 22 finished with value: 6.29 and parameters: {'iterations': 401, 'depth': 6, 'learning_rate': 0.04}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:37:34,285] Trial 23 finished with value: 6.29 and parameters: {'iterations': 476, 'depth': 6, 'learning_rate': 0.03}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:37:38,611] Trial 24 finished with value: 6.29 and parameters: {'iterations': 441, 'depth': 6, 'learning_rate': 0.04}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:37:43,183] Trial 25 finished with value: 6.29 and parameters: {'iterations': 453, 'depth': 6, 'learning_rate': 0.04}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:37:50,202] Trial 26 finished with value: 6.32 and parameters: {'iterations': 427, 'depth': 7, 'learning_rate': 0.05}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:37:52,414] Trial 27 finished with value: 6.3 and parameters: {'iterations': 404, 'depth': 5, 'learning_rate': 0.03}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:37:55,992] Trial 28 finished with value: 6.33 and parameters: {'iterations': 388, 'depth': 6, 'learning_rate': 0.05}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:37:58,859] Trial 29 finished with value: 6.33 and parameters: {'iterations': 473, 'depth': 5, 'learning_rate': 0.06}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:38:05,156] Trial 30 finished with value: 6.32 and parameters: {'iterations': 366, 'depth': 7, 'learning_rate': 0.03}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:38:09,025] Trial 31 finished with value: 6.29 and parameters: {'iterations': 430, 'depth': 6, 'learning_rate': 0.04}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:38:12,924] Trial 32 finished with value: 6.29 and parameters: {'iterations': 410, 'depth': 6, 'learning_rate': 0.04}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:38:16,942] Trial 33 finished with value: 6.29 and parameters: {'iterations': 443, 'depth': 6, 'learning_rate': 0.04}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:38:20,839] Trial 34 finished with value: 6.29 and parameters: {'iterations': 422, 'depth': 6, 'learning_rate': 0.04}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:38:25,479] Trial 35 finished with value: 6.29 and parameters: {'iterations': 448, 'depth': 6, 'learning_rate': 0.03}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:38:30,266] Trial 36 finished with value: 6.3 and parameters: {'iterations': 484, 'depth': 6, 'learning_rate': 0.04}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:38:34,086] Trial 37 finished with value: 6.34 and parameters: {'iterations': 434, 'depth': 6, 'learning_rate': 0.05}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:38:36,373] Trial 38 finished with value: 6.29 and parameters: {'iterations': 416, 'depth': 5, 'learning_rate': 0.04}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:38:39,487] Trial 39 finished with value: 6.31 and parameters: {'iterations': 348, 'depth': 6, 'learning_rate': 0.03}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:38:42,831] Trial 40 finished with value: 6.29 and parameters: {'iterations': 379, 'depth': 6, 'learning_rate': 0.04}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:38:46,472] Trial 41 finished with value: 6.29 and parameters: {'iterations': 412, 'depth': 6, 'learning_rate': 0.04}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:38:50,014] Trial 42 finished with value: 6.29 and parameters: {'iterations': 394, 'depth': 6, 'learning_rate': 0.04}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:38:54,074] Trial 43 finished with value: 6.29 and parameters: {'iterations': 419, 'depth': 6, 'learning_rate': 0.04}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:38:58,514] Trial 44 finished with value: 6.29 and parameters: {'iterations': 456, 'depth': 6, 'learning_rate': 0.03}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:39:04,745] Trial 45 finished with value: 6.29 and parameters: {'iterations': 407, 'depth': 7, 'learning_rate': 0.04}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:39:08,526] Trial 46 finished with value: 6.34 and parameters: {'iterations': 428, 'depth': 6, 'learning_rate': 0.02}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:39:12,645] Trial 47 finished with value: 6.35 and parameters: {'iterations': 467, 'depth': 6, 'learning_rate': 0.05}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:39:16,650] Trial 48 finished with value: 6.34 and parameters: {'iterations': 449, 'depth': 6, 'learning_rate': 0.05}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:39:18,460] Trial 49 finished with value: 6.36 and parameters: {'iterations': 306, 'depth': 5, 'learning_rate': 0.03}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:39:25,121] Trial 50 finished with value: 6.3 and parameters: {'iterations': 439, 'depth': 7, 'learning_rate': 0.04}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:39:28,628] Trial 51 finished with value: 6.29 and parameters: {'iterations': 399, 'depth': 6, 'learning_rate': 0.04}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:39:32,108] Trial 52 finished with value: 6.29 and parameters: {'iterations': 388, 'depth': 6, 'learning_rate': 0.04}. Best is trial 3 with value: 6.29.
[I 2023-08-24 10:39:35,413] Trial 53 finished with value: 6.28 and parameters: {'iterations': 373, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:39:38,617] Trial 54 finished with value: 6.28 and parameters: {'iterations': 368, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:39:41,797] Trial 55 finished with value: 6.28 and parameters: {'iterations': 357, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:39:44,984] Trial 56 finished with value: 6.28 and parameters: {'iterations': 361, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:39:48,126] Trial 57 finished with value: 6.31 and parameters: {'iterations': 353, 'depth': 6, 'learning_rate': 0.03}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:39:51,388] Trial 58 finished with value: 6.33 and parameters: {'iterations': 367, 'depth': 6, 'learning_rate': 0.05}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:39:54,309] Trial 59 finished with value: 6.28 and parameters: {'iterations': 331, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:39:57,416] Trial 60 finished with value: 6.28 and parameters: {'iterations': 336, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:40:00,459] Trial 61 finished with value: 6.29 and parameters: {'iterations': 329, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:40:03,408] Trial 62 finished with value: 6.29 and parameters: {'iterations': 327, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:40:06,385] Trial 63 finished with value: 6.29 and parameters: {'iterations': 338, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:40:09,209] Trial 64 finished with value: 6.29 and parameters: {'iterations': 319, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:40:12,366] Trial 65 finished with value: 6.28 and parameters: {'iterations': 357, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:40:15,557] Trial 66 finished with value: 6.28 and parameters: {'iterations': 356, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:40:18,562] Trial 67 finished with value: 6.29 and parameters: {'iterations': 342, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:40:21,898] Trial 68 finished with value: 6.3 and parameters: {'iterations': 373, 'depth': 6, 'learning_rate': 0.03}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:40:25,151] Trial 69 finished with value: 6.28 and parameters: {'iterations': 361, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:40:28,265] Trial 70 finished with value: 6.32 and parameters: {'iterations': 350, 'depth': 6, 'learning_rate': 0.05}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:40:31,410] Trial 71 finished with value: 6.28 and parameters: {'iterations': 356, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:40:34,445] Trial 72 finished with value: 6.28 and parameters: {'iterations': 360, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:40:37,751] Trial 73 finished with value: 6.28 and parameters: {'iterations': 373, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:40:40,755] Trial 74 finished with value: 6.28 and parameters: {'iterations': 337, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:40:43,613] Trial 75 finished with value: 6.29 and parameters: {'iterations': 322, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:40:46,871] Trial 76 finished with value: 6.28 and parameters: {'iterations': 367, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:40:49,977] Trial 77 finished with value: 6.29 and parameters: {'iterations': 344, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:40:53,384] Trial 78 finished with value: 6.29 and parameters: {'iterations': 381, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:40:56,554] Trial 79 finished with value: 6.28 and parameters: {'iterations': 357, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:40:59,361] Trial 80 finished with value: 6.33 and parameters: {'iterations': 313, 'depth': 6, 'learning_rate': 0.03}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:41:02,500] Trial 81 finished with value: 6.28 and parameters: {'iterations': 363, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:41:05,639] Trial 82 finished with value: 6.28 and parameters: {'iterations': 350, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:41:08,618] Trial 83 finished with value: 6.28 and parameters: {'iterations': 333, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:41:11,929] Trial 84 finished with value: 6.33 and parameters: {'iterations': 371, 'depth': 6, 'learning_rate': 0.05}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:41:15,299] Trial 85 finished with value: 6.29 and parameters: {'iterations': 378, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:41:18,343] Trial 86 finished with value: 6.29 and parameters: {'iterations': 345, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:41:21,571] Trial 87 finished with value: 6.28 and parameters: {'iterations': 362, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:41:25,704] Trial 88 finished with value: 6.29 and parameters: {'iterations': 387, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:41:29,110] Trial 89 finished with value: 6.28 and parameters: {'iterations': 356, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:41:32,145] Trial 90 finished with value: 6.28 and parameters: {'iterations': 336, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:41:35,221] Trial 91 finished with value: 6.28 and parameters: {'iterations': 358, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:41:38,353] Trial 92 finished with value: 6.28 and parameters: {'iterations': 352, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:41:41,663] Trial 93 finished with value: 6.28 and parameters: {'iterations': 369, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:41:45,113] Trial 94 finished with value: 6.29 and parameters: {'iterations': 377, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:41:48,159] Trial 95 finished with value: 6.29 and parameters: {'iterations': 342, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:41:51,401] Trial 96 finished with value: 6.28 and parameters: {'iterations': 364, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:41:54,537] Trial 97 finished with value: 6.28 and parameters: {'iterations': 355, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:41:57,495] Trial 98 finished with value: 6.29 and parameters: {'iterations': 330, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:42:00,559] Trial 99 finished with value: 6.29 and parameters: {'iterations': 348, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:42:04,097] Trial 100 finished with value: 6.32 and parameters: {'iterations': 384, 'depth': 6, 'learning_rate': 0.06}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:42:07,293] Trial 101 finished with value: 6.28 and parameters: {'iterations': 359, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:42:10,468] Trial 102 finished with value: 6.28 and parameters: {'iterations': 362, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:42:13,574] Trial 103 finished with value: 6.29 and parameters: {'iterations': 347, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:42:16,586] Trial 104 finished with value: 6.29 and parameters: {'iterations': 340, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:42:19,881] Trial 105 finished with value: 6.28 and parameters: {'iterations': 360, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:42:23,199] Trial 106 finished with value: 6.29 and parameters: {'iterations': 376, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:42:26,147] Trial 107 finished with value: 6.29 and parameters: {'iterations': 324, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:42:29,451] Trial 108 finished with value: 6.33 and parameters: {'iterations': 369, 'depth': 6, 'learning_rate': 0.05}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:42:32,563] Trial 109 finished with value: 6.28 and parameters: {'iterations': 353, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:42:35,556] Trial 110 finished with value: 6.32 and parameters: {'iterations': 334, 'depth': 6, 'learning_rate': 0.03}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:42:38,816] Trial 111 finished with value: 6.28 and parameters: {'iterations': 371, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:42:42,128] Trial 112 finished with value: 6.28 and parameters: {'iterations': 374, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:42:45,605] Trial 113 finished with value: 6.29 and parameters: {'iterations': 393, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:42:48,837] Trial 114 finished with value: 6.28 and parameters: {'iterations': 367, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:42:51,681] Trial 115 finished with value: 6.29 and parameters: {'iterations': 317, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:42:55,241] Trial 116 finished with value: 6.28 and parameters: {'iterations': 365, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:42:58,914] Trial 117 finished with value: 6.28 and parameters: {'iterations': 355, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:43:02,525] Trial 118 finished with value: 6.28 and parameters: {'iterations': 360, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:43:06,197] Trial 119 finished with value: 6.29 and parameters: {'iterations': 382, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:43:09,573] Trial 120 finished with value: 6.28 and parameters: {'iterations': 350, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:43:12,805] Trial 121 finished with value: 6.29 and parameters: {'iterations': 338, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:43:16,176] Trial 122 finished with value: 6.29 and parameters: {'iterations': 344, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:43:19,546] Trial 123 finished with value: 6.28 and parameters: {'iterations': 333, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:43:23,454] Trial 124 finished with value: 6.28 and parameters: {'iterations': 373, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:43:26,897] Trial 125 finished with value: 6.28 and parameters: {'iterations': 356, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:43:30,169] Trial 126 finished with value: 6.28 and parameters: {'iterations': 363, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:43:33,624] Trial 127 finished with value: 6.29 and parameters: {'iterations': 327, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:43:37,029] Trial 128 finished with value: 6.29 and parameters: {'iterations': 347, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:43:40,266] Trial 129 finished with value: 6.29 and parameters: {'iterations': 310, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:43:43,539] Trial 130 finished with value: 6.28 and parameters: {'iterations': 352, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:43:46,845] Trial 131 finished with value: 6.28 and parameters: {'iterations': 366, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:43:50,177] Trial 132 finished with value: 6.28 and parameters: {'iterations': 370, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:43:53,715] Trial 133 finished with value: 6.28 and parameters: {'iterations': 359, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:43:57,585] Trial 134 finished with value: 6.29 and parameters: {'iterations': 377, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:44:00,899] Trial 135 finished with value: 6.28 and parameters: {'iterations': 366, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:44:04,117] Trial 136 finished with value: 6.28 and parameters: {'iterations': 361, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:44:07,177] Trial 137 finished with value: 6.29 and parameters: {'iterations': 343, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:44:10,365] Trial 138 finished with value: 6.28 and parameters: {'iterations': 357, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:44:13,459] Trial 139 finished with value: 6.29 and parameters: {'iterations': 339, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:44:16,610] Trial 140 finished with value: 6.28 and parameters: {'iterations': 351, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:44:19,641] Trial 141 finished with value: 6.28 and parameters: {'iterations': 355, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:44:22,913] Trial 142 finished with value: 6.28 and parameters: {'iterations': 369, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:44:26,136] Trial 143 finished with value: 6.28 and parameters: {'iterations': 364, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:44:29,540] Trial 144 finished with value: 6.28 and parameters: {'iterations': 374, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:44:32,725] Trial 145 finished with value: 6.28 and parameters: {'iterations': 361, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:44:35,815] Trial 146 finished with value: 6.29 and parameters: {'iterations': 346, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:44:39,050] Trial 147 finished with value: 6.29 and parameters: {'iterations': 384, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:44:42,397] Trial 148 finished with value: 6.28 and parameters: {'iterations': 356, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:44:45,898] Trial 149 finished with value: 6.29 and parameters: {'iterations': 379, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:44:48,823] Trial 150 finished with value: 6.28 and parameters: {'iterations': 332, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:44:52,176] Trial 151 finished with value: 6.28 and parameters: {'iterations': 366, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:44:55,453] Trial 152 finished with value: 6.28 and parameters: {'iterations': 371, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:44:58,633] Trial 153 finished with value: 6.28 and parameters: {'iterations': 358, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:45:01,750] Trial 154 finished with value: 6.28 and parameters: {'iterations': 351, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:45:05,015] Trial 155 finished with value: 6.28 and parameters: {'iterations': 364, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:45:07,901] Trial 156 finished with value: 6.29 and parameters: {'iterations': 324, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:45:11,110] Trial 157 finished with value: 6.28 and parameters: {'iterations': 361, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:45:14,136] Trial 158 finished with value: 6.28 and parameters: {'iterations': 336, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:45:17,416] Trial 159 finished with value: 6.28 and parameters: {'iterations': 368, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:45:20,550] Trial 160 finished with value: 6.28 and parameters: {'iterations': 354, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:45:23,633] Trial 161 finished with value: 6.28 and parameters: {'iterations': 350, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:45:26,695] Trial 162 finished with value: 6.29 and parameters: {'iterations': 348, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:45:29,799] Trial 163 finished with value: 6.29 and parameters: {'iterations': 341, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:45:32,966] Trial 164 finished with value: 6.28 and parameters: {'iterations': 359, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:45:36,181] Trial 165 finished with value: 6.28 and parameters: {'iterations': 363, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:45:39,349] Trial 166 finished with value: 6.28 and parameters: {'iterations': 355, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:45:43,760] Trial 167 finished with value: 6.3 and parameters: {'iterations': 500, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:45:46,810] Trial 168 finished with value: 6.41 and parameters: {'iterations': 346, 'depth': 6, 'learning_rate': 0.02}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:45:49,983] Trial 169 finished with value: 6.28 and parameters: {'iterations': 353, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:45:53,267] Trial 170 finished with value: 6.28 and parameters: {'iterations': 374, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:45:56,193] Trial 171 finished with value: 6.29 and parameters: {'iterations': 329, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:45:59,190] Trial 172 finished with value: 6.28 and parameters: {'iterations': 335, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:46:02,068] Trial 173 finished with value: 6.29 and parameters: {'iterations': 319, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:46:05,070] Trial 174 finished with value: 6.29 and parameters: {'iterations': 341, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:46:08,312] Trial 175 finished with value: 6.28 and parameters: {'iterations': 362, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:46:11,575] Trial 176 finished with value: 6.28 and parameters: {'iterations': 368, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:46:14,527] Trial 177 finished with value: 6.28 and parameters: {'iterations': 332, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:46:17,694] Trial 178 finished with value: 6.28 and parameters: {'iterations': 357, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:46:20,779] Trial 179 finished with value: 6.28 and parameters: {'iterations': 350, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:46:24,047] Trial 180 finished with value: 6.29 and parameters: {'iterations': 327, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:46:27,783] Trial 181 finished with value: 6.28 and parameters: {'iterations': 362, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:46:31,081] Trial 182 finished with value: 6.28 and parameters: {'iterations': 364, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:46:34,286] Trial 183 finished with value: 6.28 and parameters: {'iterations': 358, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:46:37,613] Trial 184 finished with value: 6.28 and parameters: {'iterations': 372, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:46:40,886] Trial 185 finished with value: 6.28 and parameters: {'iterations': 367, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:46:44,170] Trial 186 finished with value: 6.28 and parameters: {'iterations': 359, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:46:47,209] Trial 187 finished with value: 6.29 and parameters: {'iterations': 338, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:46:50,331] Trial 188 finished with value: 6.28 and parameters: {'iterations': 353, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:46:53,650] Trial 189 finished with value: 6.28 and parameters: {'iterations': 370, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:46:56,726] Trial 190 finished with value: 6.29 and parameters: {'iterations': 348, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:46:59,921] Trial 191 finished with value: 6.28 and parameters: {'iterations': 357, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:47:03,245] Trial 192 finished with value: 6.28 and parameters: {'iterations': 363, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:47:06,527] Trial 193 finished with value: 6.28 and parameters: {'iterations': 354, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:47:09,432] Trial 194 finished with value: 6.29 and parameters: {'iterations': 344, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:47:12,609] Trial 195 finished with value: 6.28 and parameters: {'iterations': 360, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:47:15,869] Trial 196 finished with value: 6.28 and parameters: {'iterations': 364, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:47:19,018] Trial 197 finished with value: 6.28 and parameters: {'iterations': 356, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:47:22,142] Trial 198 finished with value: 6.28 and parameters: {'iterations': 351, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.
[I 2023-08-24 10:47:25,494] Trial 199 finished with value: 6.29 and parameters: {'iterations': 377, 'depth': 6, 'learning_rate': 0.04}. Best is trial 53 with value: 6.28.

Лучшие параметры: {'iterations': 373, 'depth': 6, 'learning_rate': 0.04}
Лучшее значение метрики: 6.28

<timed exec>:49: ExperimentalWarning:

plot_param_importances is experimental (supported from v2.2.0). The interface can change in the future.

<timed exec>:65: ExperimentalWarning:

plot_optimization_history is experimental (supported from v2.2.0). The interface can change in the future.

Wall time: 11min 36s


catboost.get_params()

{'iterations': 373,
 'learning_rate': 0.04,
 'depth': 6,
 'loss_function': 'RMSE',
 'silent': True,
 'random_state': 140823}


%%time

def objective(trial):
    # Определяем параметры для подбора
    C = trial.suggest_float('C', 0.01, 1000, step=0.01)
    epsilon = trial.suggest_float('epsilon', 0.01, 10, step=0.01)
    kernel = trial.suggest_categorical('kernel', ['linear', 'rbf', 'poly'])
    degree = trial.suggest_int('degree', 2, 5)

    # Создаем модель SVR с определенными параметрами
    model = SVR(C=C,
                epsilon=epsilon,
                kernel=kernel,
                degree=degree,
                max_iter=-1)

    # Оцениваем качество модели с помощью кросс-валидации
    score = round(abs(cross_val_score(model,
                                      features_train,
                                      target_train,
                                      cv=kf,
                                      scoring='neg_mean_absolute_error', n_jobs=-1)).mean(), 2)

    # Возвращает значение метрики
    return score.mean()

# Создаем объект Study и запускаем оптимизацию
study = optuna.create_study(direction='minimize')
study.optimize(objective, n_trials=50)

print('Лучшие параметры:', study.best_params)
print('Лучшее значение метрики:', study.best_value)

# Создаём датафрейм с значениями гиперпараметров и метрикой
results = pd.DataFrame(
    {'C': [],
     'epsilon': [],
     'kernel': [],
     'degree': [],
     'score': []}
)

# Заполняем данными из best_trials
for trial in study.best_trials:
    params = trial.params
    params['score'] = trial.value
    results = results.append(params, ignore_index=True)

# Сортируем датафрейм по убыванию значения метрики
results = results.sort_values(by='score', ascending=False)

# Выводим 10 лучших моделей
display(results.head(10))

# Создаём график важности параметров
optuna.visualization.matplotlib.plot_param_importances(study)
plt.xlabel('Importance',
           fontsize=12,
           color='DarkSlateGray')
plt.ylabel('Hyperparameters',
           fontsize=12,
           color='DarkSlateGray')
plt.suptitle('Hyperparameters Importance',
             fontsize=15,
             color='DarkSlateGray')
plt.minorticks_on()
plt.grid(which='minor',
         linestyle=':')
plt.grid(True)

#  Создаём график истории оптимизации
optuna.visualization.matplotlib.plot_optimization_history(study)
plt.tight_layout()
plt.xlabel('Iterations',
           fontsize=12,
           color='DarkSlateGray')
plt.ylabel('Objective Value',
           fontsize=12,
           color='DarkSlateGray')
plt.suptitle('Optimization History \n \n',
             fontsize=15,
             color='DarkSlateGray')
plt.minorticks_on()
plt.grid(which='minor',
         linestyle=':')
plt.grid(True)
plt.show()

[I 2023-08-24 10:47:28,072] A new study created in memory with name: no-name-b7da7403-18ec-4f79-9549-d4776d15c34a
[I 2023-08-24 10:47:28,517] Trial 0 finished with value: 7.19 and parameters: {'C': 390.02, 'epsilon': 7.71, 'kernel': 'rbf', 'degree': 4}. Best is trial 0 with value: 7.19.
[I 2023-08-24 10:47:51,516] Trial 1 finished with value: 6.55 and parameters: {'C': 772.88, 'epsilon': 5.29, 'kernel': 'linear', 'degree': 5}. Best is trial 1 with value: 6.55.
[I 2023-08-24 10:48:18,053] Trial 2 finished with value: 6.53 and parameters: {'C': 816.83, 'epsilon': 3.3499999999999996, 'kernel': 'linear', 'degree': 3}. Best is trial 2 with value: 6.53.
[I 2023-08-24 10:48:18,471] Trial 3 finished with value: 45.6 and parameters: {'C': 240.6, 'epsilon': 2.36, 'kernel': 'poly', 'degree': 5}. Best is trial 2 with value: 6.53.
[I 2023-08-24 10:48:20,773] Trial 4 finished with value: 7.68 and parameters: {'C': 703.71, 'epsilon': 0.29000000000000004, 'kernel': 'rbf', 'degree': 3}. Best is trial 2 with value: 6.53.
[I 2023-08-24 10:48:21,449] Trial 5 finished with value: 21.55 and parameters: {'C': 201.5, 'epsilon': 1.37, 'kernel': 'poly', 'degree': 4}. Best is trial 2 with value: 6.53.
[I 2023-08-24 10:48:21,616] Trial 6 finished with value: 6.94 and parameters: {'C': 101.4, 'epsilon': 8.88, 'kernel': 'rbf', 'degree': 5}. Best is trial 2 with value: 6.53.
[I 2023-08-24 10:48:22,078] Trial 7 finished with value: 8.96 and parameters: {'C': 43.35, 'epsilon': 0.72, 'kernel': 'poly', 'degree': 3}. Best is trial 2 with value: 6.53.
[I 2023-08-24 10:48:49,417] Trial 8 finished with value: 6.53 and parameters: {'C': 712.3100000000001, 'epsilon': 3.3499999999999996, 'kernel': 'linear', 'degree': 3}. Best is trial 2 with value: 6.53.
[I 2023-08-24 10:48:49,741] Trial 9 finished with value: 10.31 and parameters: {'C': 184.16, 'epsilon': 7.71, 'kernel': 'poly', 'degree': 3}. Best is trial 2 with value: 6.53.
[I 2023-08-24 10:49:17,413] Trial 10 finished with value: 6.54 and parameters: {'C': 983.48, 'epsilon': 4.6499999999999995, 'kernel': 'linear', 'degree': 2}. Best is trial 2 with value: 6.53.
[I 2023-08-24 10:49:45,131] Trial 11 finished with value: 6.53 and parameters: {'C': 657.18, 'epsilon': 3.86, 'kernel': 'linear', 'degree': 2}. Best is trial 2 with value: 6.53.
[I 2023-08-24 10:50:16,884] Trial 12 finished with value: 6.53 and parameters: {'C': 929.58, 'epsilon': 3.4, 'kernel': 'linear', 'degree': 3}. Best is trial 2 with value: 6.53.
[I 2023-08-24 10:50:30,712] Trial 13 finished with value: 6.57 and parameters: {'C': 556.52, 'epsilon': 6.11, 'kernel': 'linear', 'degree': 2}. Best is trial 2 with value: 6.53.
[I 2023-08-24 10:51:15,231] Trial 14 finished with value: 6.53 and parameters: {'C': 837.3100000000001, 'epsilon': 2.3, 'kernel': 'linear', 'degree': 4}. Best is trial 2 with value: 6.53.
[I 2023-08-24 10:51:44,450] Trial 15 finished with value: 6.52 and parameters: {'C': 477.55, 'epsilon': 3.15, 'kernel': 'linear', 'degree': 3}. Best is trial 15 with value: 6.52.
[I 2023-08-24 10:51:58,228] Trial 16 finished with value: 6.57 and parameters: {'C': 403.87, 'epsilon': 6.12, 'kernel': 'linear', 'degree': 4}. Best is trial 15 with value: 6.52.
[I 2023-08-24 10:52:16,944] Trial 17 finished with value: 6.53 and parameters: {'C': 491.12, 'epsilon': 1.69, 'kernel': 'linear', 'degree': 2}. Best is trial 15 with value: 6.52.
[I 2023-08-24 10:52:41,681] Trial 18 finished with value: 6.53 and parameters: {'C': 576.21, 'epsilon': 4.43, 'kernel': 'linear', 'degree': 3}. Best is trial 15 with value: 6.52.
[I 2023-08-24 10:52:42,585] Trial 19 finished with value: 7.15 and parameters: {'C': 326.6, 'epsilon': 2.59, 'kernel': 'rbf', 'degree': 4}. Best is trial 15 with value: 6.52.
[I 2023-08-24 10:53:08,655] Trial 20 finished with value: 6.56 and parameters: {'C': 857.09, 'epsilon': 5.9399999999999995, 'kernel': 'linear', 'degree': 2}. Best is trial 15 with value: 6.52.
[I 2023-08-24 10:53:31,921] Trial 21 finished with value: 6.53 and parameters: {'C': 669.09, 'epsilon': 3.1599999999999997, 'kernel': 'linear', 'degree': 3}. Best is trial 15 with value: 6.52.
[I 2023-08-24 10:53:56,053] Trial 22 finished with value: 6.53 and parameters: {'C': 764.3100000000001, 'epsilon': 4.0, 'kernel': 'linear', 'degree': 3}. Best is trial 15 with value: 6.52.
[I 2023-08-24 10:54:19,528] Trial 23 finished with value: 6.52 and parameters: {'C': 511.2, 'epsilon': 3.0999999999999996, 'kernel': 'linear', 'degree': 3}. Best is trial 15 with value: 6.52.
[I 2023-08-24 10:54:42,681] Trial 24 finished with value: 6.53 and parameters: {'C': 483.12, 'epsilon': 1.53, 'kernel': 'linear', 'degree': 3}. Best is trial 15 with value: 6.52.
[I 2023-08-24 10:55:02,709] Trial 25 finished with value: 6.54 and parameters: {'C': 591.68, 'epsilon': 5.09, 'kernel': 'linear', 'degree': 3}. Best is trial 15 with value: 6.52.
[I 2023-08-24 10:55:14,635] Trial 26 finished with value: 6.51 and parameters: {'C': 315.3, 'epsilon': 2.7199999999999998, 'kernel': 'linear', 'degree': 4}. Best is trial 26 with value: 6.51.
[I 2023-08-24 10:55:28,045] Trial 27 finished with value: 6.52 and parameters: {'C': 327.37, 'epsilon': 2.6199999999999997, 'kernel': 'linear', 'degree': 4}. Best is trial 26 with value: 6.51.
[I 2023-08-24 10:55:28,935] Trial 28 finished with value: 27.42 and parameters: {'C': 426.01, 'epsilon': 1.06, 'kernel': 'poly', 'degree': 4}. Best is trial 26 with value: 6.51.
[I 2023-08-24 10:55:29,143] Trial 29 finished with value: 7.24 and parameters: {'C': 282.25, 'epsilon': 9.99, 'kernel': 'rbf', 'degree': 4}. Best is trial 26 with value: 6.51.
[I 2023-08-24 10:55:30,804] Trial 30 finished with value: 7.53 and parameters: {'C': 449.3, 'epsilon': 0.08, 'kernel': 'rbf', 'degree': 5}. Best is trial 26 with value: 6.51.
[I 2023-08-24 10:55:42,558] Trial 31 finished with value: 6.51 and parameters: {'C': 336.64, 'epsilon': 2.65, 'kernel': 'linear', 'degree': 4}. Best is trial 26 with value: 6.51.
[I 2023-08-24 10:55:58,660] Trial 32 finished with value: 6.53 and parameters: {'C': 353.79, 'epsilon': 1.8900000000000001, 'kernel': 'linear', 'degree': 4}. Best is trial 26 with value: 6.51.
[I 2023-08-24 10:56:15,293] Trial 33 finished with value: 6.51 and parameters: {'C': 379.11, 'epsilon': 2.8899999999999997, 'kernel': 'linear', 'degree': 4}. Best is trial 26 with value: 6.51.
[I 2023-08-24 10:56:20,021] Trial 34 finished with value: 6.51 and parameters: {'C': 142.16, 'epsilon': 4.01, 'kernel': 'linear', 'degree': 5}. Best is trial 26 with value: 6.51.
[I 2023-08-24 10:56:27,674] Trial 35 finished with value: 6.51 and parameters: {'C': 123.89000000000001, 'epsilon': 4.01, 'kernel': 'linear', 'degree': 5}. Best is trial 26 with value: 6.51.
[I 2023-08-24 10:56:28,473] Trial 36 finished with value: 6.45 and parameters: {'C': 10.97, 'epsilon': 2.1199999999999997, 'kernel': 'linear', 'degree': 5}. Best is trial 36 with value: 6.45.
[I 2023-08-24 10:56:28,780] Trial 37 finished with value: 16.5 and parameters: {'C': 22.490000000000002, 'epsilon': 2.01, 'kernel': 'poly', 'degree': 5}. Best is trial 36 with value: 6.45.
[I 2023-08-24 10:56:42,351] Trial 38 finished with value: 6.51 and parameters: {'C': 270.34, 'epsilon': 0.8300000000000001, 'kernel': 'linear', 'degree': 5}. Best is trial 36 with value: 6.45.
[I 2023-08-24 10:56:42,995] Trial 39 finished with value: 7.01 and parameters: {'C': 200.03, 'epsilon': 2.6999999999999997, 'kernel': 'rbf', 'degree': 4}. Best is trial 36 with value: 6.45.
[I 2023-08-24 10:56:45,868] Trial 40 finished with value: 6.51 and parameters: {'C': 67.12, 'epsilon': 1.21, 'kernel': 'linear', 'degree': 4}. Best is trial 36 with value: 6.45.
[I 2023-08-24 10:56:51,618] Trial 41 finished with value: 6.51 and parameters: {'C': 141.93, 'epsilon': 3.69, 'kernel': 'linear', 'degree': 5}. Best is trial 36 with value: 6.45.
[I 2023-08-24 10:56:52,109] Trial 42 finished with value: 6.46 and parameters: {'C': 7.26, 'epsilon': 4.38, 'kernel': 'linear', 'degree': 5}. Best is trial 36 with value: 6.45.
[I 2023-08-24 10:56:54,687] Trial 43 finished with value: 6.5 and parameters: {'C': 71.9, 'epsilon': 4.59, 'kernel': 'linear', 'degree': 5}. Best is trial 36 with value: 6.45.
[I 2023-08-24 10:56:54,874] Trial 44 finished with value: 11.13 and parameters: {'C': 4.06, 'epsilon': 5.43, 'kernel': 'poly', 'degree': 5}. Best is trial 36 with value: 6.45.
[I 2023-08-24 10:56:57,240] Trial 45 finished with value: 6.57 and parameters: {'C': 96.99000000000001, 'epsilon': 6.72, 'kernel': 'linear', 'degree': 5}. Best is trial 36 with value: 6.45.
[I 2023-08-24 10:56:59,541] Trial 46 finished with value: 6.5 and parameters: {'C': 58.67, 'epsilon': 4.64, 'kernel': 'linear', 'degree': 5}. Best is trial 36 with value: 6.45.
[I 2023-08-24 10:57:02,457] Trial 47 finished with value: 6.5 and parameters: {'C': 61.91, 'epsilon': 4.68, 'kernel': 'linear', 'degree': 5}. Best is trial 36 with value: 6.45.
[I 2023-08-24 10:57:05,819] Trial 48 finished with value: 6.5 and parameters: {'C': 71.71000000000001, 'epsilon': 4.6499999999999995, 'kernel': 'linear', 'degree': 5}. Best is trial 36 with value: 6.45.
[I 2023-08-24 10:57:06,041] Trial 49 finished with value: 20.33 and parameters: {'C': 45.449999999999996, 'epsilon': 5.49, 'kernel': 'poly', 'degree': 5}. Best is trial 36 with value: 6.45.

Лучшие параметры: {'C': 10.97, 'epsilon': 2.1199999999999997, 'kernel': 'linear', 'degree': 5}
Лучшее значение метрики: 6.45

<timed exec>:54: ExperimentalWarning:

plot_param_importances is experimental (supported from v2.2.0). The interface can change in the future.

<timed exec>:70: ExperimentalWarning:

plot_optimization_history is experimental (supported from v2.2.0). The interface can change in the future.

Wall time: 9min 40s


# Создаём график контура
fig = optuna.visualization.plot_contour(study, params=["kernel", "epsilon"])
fig.show()


# Создаём график контура
fig = optuna.visualization.plot_contour(study, params=["C", "degree"])
fig.show()


%%time

def objective(trial):
    # Определяем параметры для подбора
    C = trial.suggest_float('C', 0.01, 600, step=0.01)
    epsilon = trial.suggest_float('epsilon', 1, 8, step=0.1)
    kernel = trial.suggest_categorical('kernel', ['linear'])
    degree = trial.suggest_int('degree', 3, 4)

    # Создаем модель SVR с определенными параметрами
    model = SVR(C=C,
                epsilon=epsilon,
                kernel=kernel,
                degree=degree,
                max_iter=-1)

    # Оцениваем качество модели с помощью кросс-валидации
    score = round(abs(cross_val_score(model,
                                      features_train,
                                      target_train,
                                      cv=kf,
                                      scoring='neg_mean_absolute_error', n_jobs=-1)).mean(), 2)

    # Возвращает значение метрики
    return score.mean()


# Создаем объект Study и запускаем оптимизацию
study = optuna.create_study(direction='minimize')
study.optimize(objective, n_trials=100)

print('Лучшие параметры:', study.best_params)
print('Лучшее значение метрики:', study.best_value)

# Создаём датафрейм с значениями гиперпараметров и метрикой
results = pd.DataFrame(
    {'C': [],
     'epsilon': [],
     'kernel': [],
     'degree': [],
     'score': []})

# Заполняем данными из best_trials
for trial in study.best_trials:
    params = trial.params
    params['score'] = trial.value
    results = results.append(params, ignore_index=True)

# Сортируем датафрейм по убыванию значения метрики
results = results.sort_values(by='score', ascending=False)

# Выводим 10 лучших моделей
display(results.head(10))

# Создаём график важности параметров
optuna.visualization.matplotlib.plot_param_importances(study)
plt.xlabel('Importance',
           fontsize=12,
           color='DarkSlateGray')
plt.ylabel('Hyperparameters',
           fontsize=12,
           color='DarkSlateGray')
plt.suptitle('Hyperparameters Importance',
             fontsize=15,
             color='DarkSlateGray')
plt.minorticks_on()
plt.grid(which='minor',
         linestyle=':')
plt.grid(True)

#  Создаём график истории оптимизации
optuna.visualization.matplotlib.plot_optimization_history(study)
plt.tight_layout()
plt.xlabel('Iterations',
           fontsize=12,
           color='DarkSlateGray')
plt.ylabel('Objective Value',
           fontsize=12,
           color='DarkSlateGray')
plt.suptitle('Optimization History \n \n',
             fontsize=15,
             color='DarkSlateGray')
plt.minorticks_on()
plt.grid(which='minor', linestyle=':')
plt.grid(True)
plt.show()

# Сохранение модели с наилучшими параметрами
svr = SVR(C=study.best_params['C'],
            epsilon=study.best_params['epsilon'],
            kernel=study.best_params['kernel'],
            degree=study.best_params['degree'])

[I 2023-08-24 10:57:08,292] A new study created in memory with name: no-name-a200c9c2-6db9-4536-9e7d-f570cf7c63cf
[I 2023-08-24 10:57:16,023] Trial 0 finished with value: 6.54 and parameters: {'C': 213.62, 'epsilon': 5.3, 'kernel': 'linear', 'degree': 4}. Best is trial 0 with value: 6.54.
[I 2023-08-24 10:57:29,999] Trial 1 finished with value: 6.58 and parameters: {'C': 408.39, 'epsilon': 7.1000000000000005, 'kernel': 'linear', 'degree': 4}. Best is trial 0 with value: 6.54.
[I 2023-08-24 10:57:33,419] Trial 2 finished with value: 6.57 and parameters: {'C': 143.26999999999998, 'epsilon': 6.5, 'kernel': 'linear', 'degree': 4}. Best is trial 0 with value: 6.54.
[I 2023-08-24 10:57:44,468] Trial 3 finished with value: 6.57 and parameters: {'C': 397.51, 'epsilon': 6.0, 'kernel': 'linear', 'degree': 3}. Best is trial 0 with value: 6.54.
[I 2023-08-24 10:57:45,049] Trial 4 finished with value: 6.48 and parameters: {'C': 16.700000000000003, 'epsilon': 5.6000000000000005, 'kernel': 'linear', 'degree': 3}. Best is trial 4 with value: 6.48.
[I 2023-08-24 10:57:48,349] Trial 5 finished with value: 6.54 and parameters: {'C': 125.23, 'epsilon': 5.9, 'kernel': 'linear', 'degree': 4}. Best is trial 4 with value: 6.48.
[I 2023-08-24 10:57:52,252] Trial 6 finished with value: 6.58 and parameters: {'C': 159.75, 'epsilon': 7.800000000000001, 'kernel': 'linear', 'degree': 3}. Best is trial 4 with value: 6.48.
[I 2023-08-24 10:58:10,214] Trial 7 finished with value: 6.58 and parameters: {'C': 483.48, 'epsilon': 7.4, 'kernel': 'linear', 'degree': 3}. Best is trial 4 with value: 6.48.
[I 2023-08-24 10:58:26,585] Trial 8 finished with value: 6.57 and parameters: {'C': 547.78, 'epsilon': 6.1000000000000005, 'kernel': 'linear', 'degree': 3}. Best is trial 4 with value: 6.48.
[I 2023-08-24 10:58:38,683] Trial 9 finished with value: 6.58 and parameters: {'C': 300.98, 'epsilon': 6.6000000000000005, 'kernel': 'linear', 'degree': 4}. Best is trial 4 with value: 6.48.
[I 2023-08-24 10:58:39,567] Trial 10 finished with value: 6.45 and parameters: {'C': 15.38, 'epsilon': 3.3000000000000003, 'kernel': 'linear', 'degree': 3}. Best is trial 10 with value: 6.45.
[I 2023-08-24 10:58:39,783] Trial 11 finished with value: 6.43 and parameters: {'C': 0.98, 'epsilon': 3.0, 'kernel': 'linear', 'degree': 3}. Best is trial 11 with value: 6.43.
[I 2023-08-24 10:58:40,803] Trial 12 finished with value: 6.45 and parameters: {'C': 18.970000000000002, 'epsilon': 2.8, 'kernel': 'linear', 'degree': 3}. Best is trial 11 with value: 6.43.
[I 2023-08-24 10:58:41,362] Trial 13 finished with value: 6.44 and parameters: {'C': 8.24, 'epsilon': 3.0, 'kernel': 'linear', 'degree': 3}. Best is trial 11 with value: 6.43.
[I 2023-08-24 10:58:45,599] Trial 14 finished with value: 6.52 and parameters: {'C': 103.81, 'epsilon': 1.2, 'kernel': 'linear', 'degree': 3}. Best is trial 11 with value: 6.43.
[I 2023-08-24 10:58:54,621] Trial 15 finished with value: 6.51 and parameters: {'C': 241.85999999999999, 'epsilon': 3.7, 'kernel': 'linear', 'degree': 3}. Best is trial 11 with value: 6.43.
[I 2023-08-24 10:58:58,383] Trial 16 finished with value: 6.52 and parameters: {'C': 86.05000000000001, 'epsilon': 1.7000000000000002, 'kernel': 'linear', 'degree': 3}. Best is trial 11 with value: 6.43.
[I 2023-08-24 10:59:10,885] Trial 17 finished with value: 6.52 and parameters: {'C': 299.58, 'epsilon': 2.3, 'kernel': 'linear', 'degree': 3}. Best is trial 11 with value: 6.43.
[I 2023-08-24 10:59:13,021] Trial 18 finished with value: 6.51 and parameters: {'C': 61.12, 'epsilon': 4.5, 'kernel': 'linear', 'degree': 3}. Best is trial 11 with value: 6.43.
[I 2023-08-24 10:59:20,379] Trial 19 finished with value: 6.52 and parameters: {'C': 180.63, 'epsilon': 4.2, 'kernel': 'linear', 'degree': 3}. Best is trial 11 with value: 6.43.
[I 2023-08-24 10:59:31,421] Trial 20 finished with value: 6.52 and parameters: {'C': 295.82, 'epsilon': 2.6, 'kernel': 'linear', 'degree': 4}. Best is trial 11 with value: 6.43.
[I 2023-08-24 10:59:32,049] Trial 21 finished with value: 6.44 and parameters: {'C': 10.86, 'epsilon': 3.5, 'kernel': 'linear', 'degree': 3}. Best is trial 11 with value: 6.43.
[I 2023-08-24 10:59:35,214] Trial 22 finished with value: 6.5 and parameters: {'C': 60.769999999999996, 'epsilon': 4.7, 'kernel': 'linear', 'degree': 3}. Best is trial 11 with value: 6.43.
[I 2023-08-24 10:59:35,416] Trial 23 finished with value: 6.43 and parameters: {'C': 0.89, 'epsilon': 3.2, 'kernel': 'linear', 'degree': 3}. Best is trial 11 with value: 6.43.
[I 2023-08-24 10:59:38,551] Trial 24 finished with value: 6.5 and parameters: {'C': 73.54, 'epsilon': 2.2, 'kernel': 'linear', 'degree': 3}. Best is trial 11 with value: 6.43.
[I 2023-08-24 10:59:41,629] Trial 25 finished with value: 6.49 and parameters: {'C': 63.86, 'epsilon': 3.2, 'kernel': 'linear', 'degree': 3}. Best is trial 11 with value: 6.43.
[I 2023-08-24 10:59:42,003] Trial 26 finished with value: 6.44 and parameters: {'C': 4.96, 'epsilon': 3.9000000000000004, 'kernel': 'linear', 'degree': 3}. Best is trial 11 with value: 6.43.
[I 2023-08-24 10:59:49,710] Trial 27 finished with value: 6.52 and parameters: {'C': 199.7, 'epsilon': 1.7000000000000002, 'kernel': 'linear', 'degree': 3}. Best is trial 11 with value: 6.43.
[I 2023-08-24 10:59:53,809] Trial 28 finished with value: 6.53 and parameters: {'C': 109.66000000000001, 'epsilon': 5.0, 'kernel': 'linear', 'degree': 3}. Best is trial 11 with value: 6.43.
[I 2023-08-24 11:00:02,569] Trial 29 finished with value: 6.5 and parameters: {'C': 230.49, 'epsilon': 2.8, 'kernel': 'linear', 'degree': 4}. Best is trial 11 with value: 6.43.
[I 2023-08-24 11:00:05,437] Trial 30 finished with value: 6.49 and parameters: {'C': 43.08, 'epsilon': 1.9, 'kernel': 'linear', 'degree': 3}. Best is trial 11 with value: 6.43.
[I 2023-08-24 11:00:05,736] Trial 31 finished with value: 6.44 and parameters: {'C': 2.59, 'epsilon': 3.3000000000000003, 'kernel': 'linear', 'degree': 3}. Best is trial 11 with value: 6.43.
[I 2023-08-24 11:00:07,722] Trial 32 finished with value: 6.48 and parameters: {'C': 47.32, 'epsilon': 3.7, 'kernel': 'linear', 'degree': 3}. Best is trial 11 with value: 6.43.
[I 2023-08-24 11:00:14,287] Trial 33 finished with value: 6.51 and parameters: {'C': 125.65, 'epsilon': 3.1, 'kernel': 'linear', 'degree': 3}. Best is trial 11 with value: 6.43.
[I 2023-08-24 11:00:14,472] Trial 34 finished with value: 6.44 and parameters: {'C': 0.38, 'epsilon': 3.9000000000000004, 'kernel': 'linear', 'degree': 3}. Best is trial 11 with value: 6.43.
[I 2023-08-24 11:00:35,825] Trial 35 finished with value: 6.52 and parameters: {'C': 393.56, 'epsilon': 2.3, 'kernel': 'linear', 'degree': 4}. Best is trial 11 with value: 6.43.
[I 2023-08-24 11:00:37,872] Trial 36 finished with value: 6.49 and parameters: {'C': 40.74, 'epsilon': 4.300000000000001, 'kernel': 'linear', 'degree': 3}. Best is trial 11 with value: 6.43.
[I 2023-08-24 11:00:44,766] Trial 37 finished with value: 6.52 and parameters: {'C': 151.73999999999998, 'epsilon': 1.2, 'kernel': 'linear', 'degree': 3}. Best is trial 11 with value: 6.43.
[I 2023-08-24 11:00:48,756] Trial 38 finished with value: 6.52 and parameters: {'C': 94.69000000000001, 'epsilon': 5.1000000000000005, 'kernel': 'linear', 'degree': 3}. Best is trial 11 with value: 6.43.
[I 2023-08-24 11:01:13,048] Trial 39 finished with value: 6.51 and parameters: {'C': 596.4, 'epsilon': 2.8, 'kernel': 'linear', 'degree': 4}. Best is trial 11 with value: 6.43.
[I 2023-08-24 11:01:28,856] Trial 40 finished with value: 6.52 and parameters: {'C': 445.12, 'epsilon': 3.6, 'kernel': 'linear', 'degree': 3}. Best is trial 11 with value: 6.43.
[I 2023-08-24 11:01:30,213] Trial 41 finished with value: 6.47 and parameters: {'C': 27.930000000000003, 'epsilon': 4.0, 'kernel': 'linear', 'degree': 3}. Best is trial 11 with value: 6.43.
[I 2023-08-24 11:01:30,592] Trial 42 finished with value: 6.43 and parameters: {'C': 5.5, 'epsilon': 3.5, 'kernel': 'linear', 'degree': 3}. Best is trial 11 with value: 6.43.
[I 2023-08-24 11:01:32,373] Trial 43 finished with value: 6.47 and parameters: {'C': 44.9, 'epsilon': 3.5, 'kernel': 'linear', 'degree': 3}. Best is trial 11 with value: 6.43.
[I 2023-08-24 11:01:37,871] Trial 44 finished with value: 6.51 and parameters: {'C': 128.34, 'epsilon': 3.0, 'kernel': 'linear', 'degree': 3}. Best is trial 11 with value: 6.43.
[I 2023-08-24 11:01:41,909] Trial 45 finished with value: 6.5 and parameters: {'C': 91.44000000000001, 'epsilon': 2.5, 'kernel': 'linear', 'degree': 3}. Best is trial 11 with value: 6.43.
[I 2023-08-24 11:01:42,981] Trial 46 finished with value: 6.46 and parameters: {'C': 23.990000000000002, 'epsilon': 3.4000000000000004, 'kernel': 'linear', 'degree': 3}. Best is trial 11 with value: 6.43.
[I 2023-08-24 11:01:56,346] Trial 47 finished with value: 6.53 and parameters: {'C': 358.94, 'epsilon': 4.7, 'kernel': 'linear', 'degree': 3}. Best is trial 11 with value: 6.43.
[I 2023-08-24 11:02:00,071] Trial 48 finished with value: 6.49 and parameters: {'C': 76.41000000000001, 'epsilon': 3.0, 'kernel': 'linear', 'degree': 3}. Best is trial 11 with value: 6.43.
[I 2023-08-24 11:02:01,723] Trial 49 finished with value: 6.47 and parameters: {'C': 30.67, 'epsilon': 2.1, 'kernel': 'linear', 'degree': 3}. Best is trial 11 with value: 6.43.
[I 2023-08-24 11:02:05,016] Trial 50 finished with value: 6.53 and parameters: {'C': 108.22000000000001, 'epsilon': 5.6000000000000005, 'kernel': 'linear', 'degree': 3}. Best is trial 11 with value: 6.43.
[I 2023-08-24 11:02:05,459] Trial 51 finished with value: 6.44 and parameters: {'C': 6.91, 'epsilon': 3.8000000000000003, 'kernel': 'linear', 'degree': 3}. Best is trial 11 with value: 6.43.
[I 2023-08-24 11:02:05,637] Trial 52 finished with value: 6.45 and parameters: {'C': 0.54, 'epsilon': 4.2, 'kernel': 'linear', 'degree': 3}. Best is trial 11 with value: 6.43.
[I 2023-08-24 11:02:07,094] Trial 53 finished with value: 6.46 and parameters: {'C': 27.500000000000004, 'epsilon': 2.6, 'kernel': 'linear', 'degree': 3}. Best is trial 11 with value: 6.43.
[I 2023-08-24 11:02:10,084] Trial 54 finished with value: 6.49 and parameters: {'C': 61.48, 'epsilon': 4.0, 'kernel': 'linear', 'degree': 3}. Best is trial 11 with value: 6.43.
[I 2023-08-24 11:02:11,146] Trial 55 finished with value: 6.45 and parameters: {'C': 23.290000000000003, 'epsilon': 3.5, 'kernel': 'linear', 'degree': 3}. Best is trial 11 with value: 6.43.
[I 2023-08-24 11:02:14,766] Trial 56 finished with value: 6.49 and parameters: {'C': 79.26, 'epsilon': 3.0, 'kernel': 'linear', 'degree': 3}. Best is trial 11 with value: 6.43.
[I 2023-08-24 11:02:23,175] Trial 57 finished with value: 6.52 and parameters: {'C': 175.94, 'epsilon': 4.4, 'kernel': 'linear', 'degree': 3}. Best is trial 11 with value: 6.43.
[I 2023-08-24 11:02:32,768] Trial 58 finished with value: 6.53 and parameters: {'C': 268.36, 'epsilon': 4.7, 'kernel': 'linear', 'degree': 3}. Best is trial 11 with value: 6.43.
[I 2023-08-24 11:02:35,611] Trial 59 finished with value: 6.49 and parameters: {'C': 55.51, 'epsilon': 3.2, 'kernel': 'linear', 'degree': 3}. Best is trial 11 with value: 6.43.
[I 2023-08-24 11:02:35,804] Trial 60 finished with value: 6.42 and parameters: {'C': 0.3, 'epsilon': 2.5, 'kernel': 'linear', 'degree': 3}. Best is trial 60 with value: 6.42.
[I 2023-08-24 11:02:36,010] Trial 61 finished with value: 6.43 and parameters: {'C': 0.59, 'epsilon': 2.6, 'kernel': 'linear', 'degree': 3}. Best is trial 60 with value: 6.42.
[I 2023-08-24 11:02:37,839] Trial 62 finished with value: 6.48 and parameters: {'C': 40.54, 'epsilon': 2.0, 'kernel': 'linear', 'degree': 3}. Best is trial 60 with value: 6.42.
[I 2023-08-24 11:02:38,797] Trial 63 finished with value: 6.46 and parameters: {'C': 17.05, 'epsilon': 2.5, 'kernel': 'linear', 'degree': 3}. Best is trial 60 with value: 6.42.
[I 2023-08-24 11:02:38,977] Trial 64 finished with value: 6.46 and parameters: {'C': 0.09, 'epsilon': 2.6, 'kernel': 'linear', 'degree': 3}. Best is trial 60 with value: 6.42.
[I 2023-08-24 11:02:42,634] Trial 65 finished with value: 6.51 and parameters: {'C': 68.4, 'epsilon': 1.6, 'kernel': 'linear', 'degree': 3}. Best is trial 60 with value: 6.42.
[I 2023-08-24 11:02:45,039] Trial 66 finished with value: 6.47 and parameters: {'C': 42.17, 'epsilon': 2.8, 'kernel': 'linear', 'degree': 3}. Best is trial 60 with value: 6.42.
[I 2023-08-24 11:02:45,631] Trial 67 finished with value: 6.54 and parameters: {'C': 21.23, 'epsilon': 6.800000000000001, 'kernel': 'linear', 'degree': 3}. Best is trial 60 with value: 6.42.
[I 2023-08-24 11:02:48,146] Trial 68 finished with value: 6.48 and parameters: {'C': 53.089999999999996, 'epsilon': 2.3, 'kernel': 'linear', 'degree': 4}. Best is trial 60 with value: 6.42.
[I 2023-08-24 11:02:48,910] Trial 69 finished with value: 6.45 and parameters: {'C': 16.610000000000003, 'epsilon': 3.3000000000000003, 'kernel': 'linear', 'degree': 3}. Best is trial 60 with value: 6.42.
[I 2023-08-24 11:03:14,086] Trial 70 finished with value: 6.53 and parameters: {'C': 514.25, 'epsilon': 1.8, 'kernel': 'linear', 'degree': 3}. Best is trial 60 with value: 6.42.
[I 2023-08-24 11:03:14,889] Trial 71 finished with value: 6.45 and parameters: {'C': 12.81, 'epsilon': 3.7, 'kernel': 'linear', 'degree': 3}. Best is trial 60 with value: 6.42.
[I 2023-08-24 11:03:16,815] Trial 72 finished with value: 6.49 and parameters: {'C': 39.44, 'epsilon': 1.4, 'kernel': 'linear', 'degree': 3}. Best is trial 60 with value: 6.42.
[I 2023-08-24 11:03:20,567] Trial 73 finished with value: 6.5 and parameters: {'C': 78.82000000000001, 'epsilon': 3.1, 'kernel': 'linear', 'degree': 3}. Best is trial 60 with value: 6.42.
[I 2023-08-24 11:03:20,988] Trial 74 finished with value: 6.44 and parameters: {'C': 5.46, 'epsilon': 2.8, 'kernel': 'linear', 'degree': 3}. Best is trial 60 with value: 6.42.
[I 2023-08-24 11:03:23,718] Trial 75 finished with value: 6.49 and parameters: {'C': 60.61, 'epsilon': 4.0, 'kernel': 'linear', 'degree': 3}. Best is trial 60 with value: 6.42.
[I 2023-08-24 11:03:27,379] Trial 76 finished with value: 6.5 and parameters: {'C': 94.49000000000001, 'epsilon': 3.4000000000000004, 'kernel': 'linear', 'degree': 3}. Best is trial 60 with value: 6.42.
[I 2023-08-24 11:03:28,914] Trial 77 finished with value: 6.47 and parameters: {'C': 32.17, 'epsilon': 2.4000000000000004, 'kernel': 'linear', 'degree': 3}. Best is trial 60 with value: 6.42.
[I 2023-08-24 11:03:33,357] Trial 78 finished with value: 6.5 and parameters: {'C': 113.4, 'epsilon': 2.9000000000000004, 'kernel': 'linear', 'degree': 3}. Best is trial 60 with value: 6.42.
[I 2023-08-24 11:03:45,982] Trial 79 finished with value: 6.51 and parameters: {'C': 338.51, 'epsilon': 3.5, 'kernel': 'linear', 'degree': 3}. Best is trial 60 with value: 6.42.
[I 2023-08-24 11:03:51,800] Trial 80 finished with value: 6.5 and parameters: {'C': 133.75, 'epsilon': 2.7, 'kernel': 'linear', 'degree': 3}. Best is trial 60 with value: 6.42.
[I 2023-08-24 11:03:52,075] Trial 81 finished with value: 6.44 and parameters: {'C': 1.42, 'epsilon': 3.2, 'kernel': 'linear', 'degree': 3}. Best is trial 60 with value: 6.42.
[I 2023-08-24 11:03:53,306] Trial 82 finished with value: 6.45 and parameters: {'C': 18.42, 'epsilon': 3.8000000000000003, 'kernel': 'linear', 'degree': 3}. Best is trial 60 with value: 6.42.
[I 2023-08-24 11:03:55,232] Trial 83 finished with value: 6.47 and parameters: {'C': 34.97, 'epsilon': 3.3000000000000003, 'kernel': 'linear', 'degree': 3}. Best is trial 60 with value: 6.42.
[I 2023-08-24 11:03:56,778] Trial 84 finished with value: 6.58 and parameters: {'C': 52.72, 'epsilon': 8.0, 'kernel': 'linear', 'degree': 3}. Best is trial 60 with value: 6.42.
[I 2023-08-24 11:03:57,619] Trial 85 finished with value: 6.45 and parameters: {'C': 14.65, 'epsilon': 3.6, 'kernel': 'linear', 'degree': 3}. Best is trial 60 with value: 6.42.
[I 2023-08-24 11:04:00,624] Trial 86 finished with value: 6.46 and parameters: {'C': 31.450000000000003, 'epsilon': 3.0, 'kernel': 'linear', 'degree': 3}. Best is trial 60 with value: 6.42.
[I 2023-08-24 11:04:00,859] Trial 87 finished with value: 6.44 and parameters: {'C': 1.99, 'epsilon': 4.2, 'kernel': 'linear', 'degree': 3}. Best is trial 60 with value: 6.42.
[I 2023-08-24 11:04:03,758] Trial 88 finished with value: 6.49 and parameters: {'C': 71.80000000000001, 'epsilon': 3.4000000000000004, 'kernel': 'linear', 'degree': 3}. Best is trial 60 with value: 6.42.
[I 2023-08-24 11:04:05,563] Trial 89 finished with value: 6.48 and parameters: {'C': 44.04, 'epsilon': 3.9000000000000004, 'kernel': 'linear', 'degree': 3}. Best is trial 60 with value: 6.42.
[I 2023-08-24 11:04:06,558] Trial 90 finished with value: 6.45 and parameters: {'C': 15.0, 'epsilon': 2.2, 'kernel': 'linear', 'degree': 3}. Best is trial 60 with value: 6.42.
[I 2023-08-24 11:04:07,795] Trial 91 finished with value: 6.46 and parameters: {'C': 27.060000000000002, 'epsilon': 3.7, 'kernel': 'linear', 'degree': 3}. Best is trial 60 with value: 6.42.
[I 2023-08-24 11:04:08,614] Trial 92 finished with value: 6.46 and parameters: {'C': 11.58, 'epsilon': 4.1, 'kernel': 'linear', 'degree': 3}. Best is trial 60 with value: 6.42.
[I 2023-08-24 11:04:08,842] Trial 93 finished with value: 6.43 and parameters: {'C': 0.78, 'epsilon': 3.1, 'kernel': 'linear', 'degree': 3}. Best is trial 60 with value: 6.42.
[I 2023-08-24 11:04:11,551] Trial 94 finished with value: 6.48 and parameters: {'C': 51.54, 'epsilon': 3.1, 'kernel': 'linear', 'degree': 3}. Best is trial 60 with value: 6.42.
[I 2023-08-24 11:04:12,825] Trial 95 finished with value: 6.46 and parameters: {'C': 29.51, 'epsilon': 2.6, 'kernel': 'linear', 'degree': 3}. Best is trial 60 with value: 6.42.
[I 2023-08-24 11:04:13,047] Trial 96 finished with value: 6.43 and parameters: {'C': 0.68, 'epsilon': 2.9000000000000004, 'kernel': 'linear', 'degree': 3}. Best is trial 60 with value: 6.42.
[I 2023-08-24 11:04:13,279] Trial 97 finished with value: 6.43 and parameters: {'C': 0.98, 'epsilon': 2.9000000000000004, 'kernel': 'linear', 'degree': 3}. Best is trial 60 with value: 6.42.
[I 2023-08-24 11:04:14,183] Trial 98 finished with value: 6.45 and parameters: {'C': 13.95, 'epsilon': 2.9000000000000004, 'kernel': 'linear', 'degree': 4}. Best is trial 60 with value: 6.42.
[I 2023-08-24 11:04:17,964] Trial 99 finished with value: 6.5 and parameters: {'C': 86.69000000000001, 'epsilon': 2.5, 'kernel': 'linear', 'degree': 3}. Best is trial 60 with value: 6.42.

Лучшие параметры: {'C': 0.3, 'epsilon': 2.5, 'kernel': 'linear', 'degree': 3}
Лучшее значение метрики: 6.42

<timed exec>:54: ExperimentalWarning:

plot_param_importances is experimental (supported from v2.2.0). The interface can change in the future.

<timed exec>:70: ExperimentalWarning:

plot_optimization_history is experimental (supported from v2.2.0). The interface can change in the future.

Wall time: 7min 12s

svr

SVR(C=0.3, epsilon=2.5, kernel='linear')


def calculate_metric(model):
    # Вычисление средней абсолютной ошибки с помощью кросс-валидации
    scores = cross_val_score(model,
                             features_train,
                             target_train,
                             cv=kf,
                             scoring='neg_mean_absolute_error', n_jobs=-1)
    metric = round(abs(scores.mean()), 2)
    
    return metric

metrics = {}
metrics['LinearRegression'] = calculate_metric(regressor)
metrics['RandomForestRegressor'] = calculate_metric(forest)
metrics['LGBMRegressor'] = calculate_metric(lgbm)
metrics['CatBoostRegressor'] = calculate_metric(catboost)
metrics['SVR'] = calculate_metric(svr)

models = list(metrics.keys())
values = list(metrics.values())
colors = ['blue', 'green', 'red', 'purple', 'yellow']

# Построение графика сравнения метрик
plt.bar(models, values, color=colors)
plt.xlabel('Модель',
           fontsize=12,
           color='DarkSlateGray')
plt.ylabel('Метрика',
           fontsize=12,
           color='DarkSlateGray')
plt.title('Сравнение метрик',
          fontsize=15,
          color='DarkSlateGray')
plt.xticks(rotation=90)
for i, v in enumerate(values):
    plt.text(i, v, str(v), ha='center', va='bottom')

plt.show()


model = DummyRegressor()
model.fit(features_train, target_train)
lgbm_mae_test = round(mae(target_test, model.predict(features_test)), 2)
print('На простой модели регрессии МАЕ = ', lgbm_mae_test)

На простой модели регрессии МАЕ =  8.69


lgbm.fit(features_train, target_train)
lgbm_mae_test = round(mae(target_test, lgbm.predict(features_test)), 2)
print('На тестовой выборке МАЕ = ', lgbm_mae_test)

На тестовой выборке МАЕ =  6.62


lgbm

LGBMRegressor(colsample_bytree=0.7, learning_rate=0.09, max_depth=2,
              min_child_samples=7, n_estimators=276, random_state=140823,
              subsample=0.76)


importance = lgbm.feature_importances_
# Сортируем индексы важности модели в порядке убывания
sorted_indices = np.argsort(importance)[::-1] 
# Получаем отсортированные названия признаков
sorted_features = features_train.columns[sorted_indices] 
 # Получаем отсортированные значения важности признаков
sorted_importance = importance[sorted_indices]
# Вычисляем общую важность всех признаков
total_importance = np.sum(sorted_importance)
 # Вычисляем процентную важность каждого признака
sorted_importance_percent = (sorted_importance / total_importance) * 100  

# Создаем график
plt.figure(figsize=(15, 5))
 # Строим столбчатую диаграмму с отсортированными признаками и их важностью
plt.bar(sorted_features, sorted_importance_percent)  
plt.xticks(rotation='vertical')  
plt.xlabel('Признаки',
           fontsize=12,
           color='DarkSlateGray')  
plt.ylabel('Важность (%)',
           fontsize=12,
           color='DarkSlateGray')  
plt.title('Важность признаков',
          fontsize=15,
          color='DarkSlateGray')  
plt.minorticks_on() 
plt.grid(which='minor',
         linestyle=':')  
plt.grid(True) 
plt.show()

print("Таблица значимости признаков:")
df_importance = pd.DataFrame({"Признаки": sorted_features,
                              "Важность (%)": sorted_importance_percent}).style.background_gradient('coolwarm')
df_importance

Таблица значимости признаков:

	key	Начало нагрева дугой	Конец нагрева дугой	Активная мощность	Реактивная мощность
0	1	2019-05-03 11:02:14	2019-05-03 11:06:02	0.305130	0.211253
1	1	2019-05-03 11:07:28	2019-05-03 11:10:33	0.765658	0.477438
2	1	2019-05-03 11:11:44	2019-05-03 11:14:36	0.580313	0.430460
3	1	2019-05-03 11:18:14	2019-05-03 11:24:19	0.518496	0.379979
4	1	2019-05-03 11:26:09	2019-05-03 11:28:37	0.867133	0.643691
...	...	...	...	...	...
14871	3241	2019-09-06 16:49:05	2019-09-06 16:51:42	0.439735	0.299579
14872	3241	2019-09-06 16:55:11	2019-09-06 16:58:11	0.646498	0.458240
14873	3241	2019-09-06 17:06:48	2019-09-06 17:09:52	1.039726	0.769302
14874	3241	2019-09-06 17:21:58	2019-09-06 17:22:55	0.530267	0.361543
14875	3241	2019-09-06 17:24:54	2019-09-06 17:26:15	0.389057	0.251347

	key	Газ 1
0	1	29.749986
1	2	12.555561
2	3	28.554793
3	4	18.841219
4	5	5.413692
...	...	...
3234	3237	5.543905
3235	3238	6.745669
3236	3239	16.023518
3237	3240	11.863103
3238	3241	12.680959

	key	Начало нагрева дугой	Конец нагрева дугой	Активная мощность	Реактивная мощность
0	1	2019-05-03 11:02:14	2019-05-03 11:06:02	0.305130	0.211253
1	1	2019-05-03 11:07:28	2019-05-03 11:10:33	0.765658	0.477438
2	1	2019-05-03 11:11:44	2019-05-03 11:14:36	0.580313	0.430460
3	1	2019-05-03 11:18:14	2019-05-03 11:24:19	0.518496	0.379979
4	1	2019-05-03 11:26:09	2019-05-03 11:28:37	0.867133	0.643691

	key	Время замера	Температура
0	1	2019-05-03 11:02:04	1571.0
1	1	2019-05-03 11:07:18	1604.0
2	1	2019-05-03 11:11:34	1618.0
3	1	2019-05-03 11:18:04	1601.0
4	1	2019-05-03 11:25:59	1606.0

	key	Начало нагрева дугой	Конец нагрева дугой	Активная мощность	Реактивная мощность
count	14876.000000	14876	14876	14876.000000	14876.000000
mean	1615.220422	2019-07-05 12:25:51.921081088	2019-07-05 12:28:43.592027392	0.662752	0.438986
min	1.000000	2019-05-03 11:02:14	2019-05-03 11:06:02	0.223120	-715.479924
25%	806.000000	2019-06-03 23:18:23.249999872	2019-06-03 23:21:35	0.467115	0.337175
50%	1617.000000	2019-07-03 01:31:26.500000	2019-07-03 01:35:13	0.599587	0.441639
75%	2429.000000	2019-08-07 22:52:20.750000128	2019-08-07 22:56:47	0.830070	0.608201
max	3241.000000	2019-09-06 17:24:54	2019-09-06 17:26:15	1.463773	1.270284
std	934.571502	NaN	NaN	0.258885	5.873485

Содержание.

Библиотеки и константы.¶

Знакомство с данными.¶

Предобработка данных.¶

`df_arc` — данные об электродах.¶

`df_bulk` — данные о подаче сыпучих материалов (объём)¶

`df_bulk_time` — данные о подаче сыпучих материалов (время).¶

`df_gas` — данные о продувке сплава газом¶

`df_temp` — результаты измерения температуры.¶

`df_wire` — данные о проволочных материалах (объём).¶

`df_wire_time` — данные о проволочных материалах (время).¶

Моменты, которые надо уточнить и согласовать с представителем заказчика.¶

Результат предобработки.¶

Объединение таблиц.¶

Обучение моделей.¶

Подготовка выборок.¶

Вычислим среднюю абсолютную ошибку (MAE) для модели, которая всегда предсказывает среднее значение целевой переменной.¶

Линейная регрессия.¶

Случайный лес RandomForestRegressor.¶

Градиентный бустинг LightGBM¶

Градиентный бустинг CatBoostRegressor.¶

Support Vector Regression.¶

Проанализируем качество моделей.¶

Тестирование лучшей модели.¶

Степень влияния признаков на целевой показатель.¶

Отчёт по решению.¶

	key	Bulk 1	Bulk 2	Bulk 3	Bulk 4	Bulk 5	Bulk 6	Bulk 7	Bulk 8	Bulk 9	Bulk 10	Bulk 11	Bulk 12	Bulk 13	Bulk 14	Bulk 15
0	1	NaN	NaN	NaN	43.0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	206.0	NaN	150.0	154.0
1	2	NaN	NaN	NaN	73.0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	206.0	NaN	149.0	154.0
2	3	NaN	NaN	NaN	34.0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	205.0	NaN	152.0	153.0
3	4	NaN	NaN	NaN	81.0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	207.0	NaN	153.0	154.0
4	5	NaN	NaN	NaN	78.0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	203.0	NaN	151.0	152.0
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
3124	3237	NaN	NaN	170.0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	252.0	NaN	130.0	206.0
3125	3238	NaN	NaN	126.0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	254.0	NaN	108.0	106.0
3126	3239	NaN	NaN	NaN	NaN	NaN	114.0	NaN	NaN	NaN	NaN	NaN	158.0	NaN	270.0	88.0
3127	3240	NaN	NaN	NaN	NaN	NaN	26.0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	192.0	54.0
3128	3241	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	180.0	52.0

	key	Bulk 1	Bulk 2	Bulk 3	Bulk 4	Bulk 5	Bulk 6	Bulk 7	Bulk 8	Bulk 9	Bulk 10	Bulk 11	Bulk 12	Bulk 13	Bulk 14	Bulk 15
0	1	NaN	NaN	NaN	2019-05-03 11:28:48	NaN	NaN	NaN	NaN	NaN	NaN	NaN	2019-05-03 11:24:31	NaN	2019-05-03 11:14:50	2019-05-03 11:10:43
1	2	NaN	NaN	NaN	2019-05-03 11:36:50	NaN	NaN	NaN	NaN	NaN	NaN	NaN	2019-05-03 11:53:30	NaN	2019-05-03 11:48:37	2019-05-03 11:44:39
2	3	NaN	NaN	NaN	2019-05-03 12:32:39	NaN	NaN	NaN	NaN	NaN	NaN	NaN	2019-05-03 12:27:13	NaN	2019-05-03 12:21:01	2019-05-03 12:16:16
3	4	NaN	NaN	NaN	2019-05-03 12:43:22	NaN	NaN	NaN	NaN	NaN	NaN	NaN	2019-05-03 12:58:00	NaN	2019-05-03 12:51:11	2019-05-03 12:46:36
4	5	NaN	NaN	NaN	2019-05-03 13:30:47	NaN	NaN	NaN	NaN	NaN	NaN	NaN	2019-05-03 13:30:47	NaN	2019-05-03 13:34:12	2019-05-03 13:30:47
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
3124	3237	NaN	NaN	2019-09-06 11:54:15	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	2019-09-06 11:49:45	NaN	2019-09-06 11:45:22	2019-09-06 11:40:06
3125	3238	NaN	NaN	2019-09-06 12:26:52	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	2019-09-06 12:18:35	NaN	2019-09-06 12:31:49	2019-09-06 12:26:52
3126	3239	NaN	NaN	NaN	NaN	NaN	2019-09-06 15:06:00	NaN	NaN	NaN	NaN	NaN	2019-09-06 15:01:44	NaN	2019-09-06 14:58:15	2019-09-06 14:48:06
3127	3240	NaN	NaN	NaN	NaN	NaN	2019-09-06 16:24:28	NaN	NaN	NaN	NaN	NaN	NaN	NaN	2019-09-06 16:07:29	2019-09-06 16:01:34
3128	3241	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	2019-09-06 17:26:33	2019-09-06 17:23:15

	key	Wire 1	Wire 2	Wire 3	Wire 4	Wire 5	Wire 6	Wire 7	Wire 8	Wire 9
0	1	60.059998	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
1	2	96.052315	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
2	3	91.160157	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
3	4	89.063515	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
4	5	89.238236	9.11456	NaN	NaN	NaN	NaN	NaN	NaN	NaN
...	...	...	...	...	...	...	...	...	...	...
3076	3237	38.088959	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
3077	3238	56.128799	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
3078	3239	143.357761	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
3079	3240	34.070400	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
3080	3241	63.117595	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN

	key	Wire 1	Wire 2	Wire 3	Wire 4	Wire 5	Wire 6	Wire 7	Wire 8	Wire 9
0	1	2019-05-03 11:06:19	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
1	2	2019-05-03 11:36:50	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
2	3	2019-05-03 12:11:46	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
3	4	2019-05-03 12:43:22	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
4	5	2019-05-03 13:20:44	2019-05-03 13:15:34	NaN	NaN	NaN	NaN	NaN	NaN	NaN
...	...	...	...	...	...	...	...	...	...	...
3076	3237	2019-09-06 11:33:38	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
3077	3238	2019-09-06 12:18:35	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
3078	3239	2019-09-06 14:36:11	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
3079	3240	2019-09-06 15:33:55	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
3080	3241	2019-09-06 17:10:06	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN

	кол-во обработок	кол-во партий	% кол-ва партий
0	4	892	27.753578
1	5	759	23.615432
2	3	520	16.179216
3	6	490	15.245800
4	7	205	6.378345
5	2	174	5.413815
6	8	84	2.613566
7	1	39	1.213441
8	9	28	0.871189
9	10	9	0.280025
10	11	5	0.155569
11	12	3	0.093342
12	13	2	0.062228
13	15	2	0.062228
14	14	1	0.031114
15	16	1	0.031114

	Активная мощность	Реактивная мощность
Активная мощность	1.00	0.97
Реактивная мощность	0.97	1.00

	количество этапов обработки	Начало нагрева дугой	Конец нагрева дугой	Активная мощность	Реактивная мощность	Полная мощность	Продолжительность дуги	Потребляемая энергия
key
1	5	2019-05-03 11:02:14	2019-05-03 11:28:37	3.036730	2.142821	3.718736	1098.0	770.282114
2	4	2019-05-03 11:34:14	2019-05-03 11:53:18	2.139408	1.453357	2.588349	811.0	481.760005
3	5	2019-05-03 12:06:54	2019-05-03 12:32:19	4.063641	2.937457	5.019223	655.0	722.837668
4	4	2019-05-03 12:39:37	2019-05-03 12:57:50	2.706489	2.056992	3.400038	741.0	683.455597
5	4	2019-05-03 13:11:13	2019-05-03 13:33:55	2.252950	1.687991	2.816980	869.0	512.169934
...	...	...	...	...	...	...	...	...
3237	5	2019-09-06 11:31:25	2019-09-06 11:53:55	2.541872	2.025417	3.250657	909.0	630.503534
3238	3	2019-09-06 12:16:25	2019-09-06 12:31:35	1.374821	1.038103	1.723937	546.0	286.052252
3239	8	2019-09-06 14:17:00	2019-09-06 15:05:50	4.848005	3.541541	6.014480	1216.0	941.538764
3240	5	2019-09-06 15:25:31	2019-09-06 16:24:15	3.317679	2.373552	4.082920	839.0	657.439848
3241	5	2019-09-06 16:49:05	2019-09-06 17:26:15	3.045283	2.140011	3.722880	659.0	538.258300

	Bulk 1	Bulk 2	Bulk 3	Bulk 4	Bulk 5	Bulk 6	Bulk 7	Bulk 8	Bulk 9	Bulk 10	Bulk 11	Bulk 12	Bulk 13	Bulk 14	Bulk 15
count	252.000000	22.000000	1298.000000	1014.000000	77.000000	576.000000	25.000000	1.0	19.000000	176.000000	177.000000	2450.000000	18.000000	2806.000000	2248.000000
mean	39.242063	253.045455	113.879045	104.394477	107.025974	118.925347	305.600000	49.0	76.315789	83.284091	76.819209	260.471020	181.111111	170.284747	160.513345
std	18.277654	21.180578	75.483494	48.184126	81.790646	72.057776	191.022904	NaN	21.720581	26.060347	59.655365	120.649269	46.088009	65.868652	51.765319
min	10.000000	228.000000	6.000000	12.000000	11.000000	17.000000	47.000000	49.0	63.000000	24.000000	8.000000	53.000000	151.000000	16.000000	1.000000
25%	27.000000	242.000000	58.000000	72.000000	70.000000	69.750000	155.000000	49.0	66.000000	64.000000	25.000000	204.000000	153.250000	119.000000	105.000000
50%	31.000000	251.500000	97.500000	102.000000	86.000000	100.000000	298.000000	49.0	68.000000	86.500000	64.000000	208.000000	155.500000	151.000000	160.000000
75%	46.000000	257.750000	152.000000	133.000000	132.000000	157.000000	406.000000	49.0	70.500000	102.000000	106.000000	316.000000	203.500000	205.750000	205.000000
max	185.000000	325.000000	454.000000	281.000000	603.000000	503.000000	772.000000	49.0	147.000000	159.000000	313.000000	1849.000000	305.000000	636.000000	405.000000

	start_time_bulk	finish_time_bulk	duration_bulk
key
1	2019-05-03 11:10:43	2019-05-03 11:28:48	1085.0
2	2019-05-03 11:36:50	2019-05-03 11:53:30	1000.0
3	2019-05-03 12:16:16	2019-05-03 12:32:39	983.0
4	2019-05-03 12:43:22	2019-05-03 12:58:00	878.0
5	2019-05-03 13:30:47	2019-05-03 13:34:12	205.0
...	...	...	...
3237	2019-09-06 11:40:06	2019-09-06 11:54:15	849.0
3238	2019-09-06 12:18:35	2019-09-06 12:31:49	794.0
3239	2019-09-06 14:48:06	2019-09-06 15:06:00	1074.0
3240	2019-09-06 16:01:34	2019-09-06 16:24:28	1374.0
3241	2019-09-06 17:23:15	2019-09-06 17:26:33	198.0

	start_time_bulk	finish_time_bulk	duration_bulk
count	3129	3129	3129.000000
mean	2019-07-05 20:31:19.313518336	2019-07-05 20:47:24.220198144	964.906679
min	2019-05-03 11:10:43	2019-05-03 11:28:48	0.000000
25%	2019-06-04 09:12:47	2019-06-04 09:22:40	485.000000
50%	2019-07-03 04:58:44	2019-07-03 05:17:25	877.000000
75%	2019-08-08 00:03:27	2019-08-08 00:23:14	1311.000000
max	2019-09-06 17:23:15	2019-09-06 17:26:33	13683.000000
std	NaN	NaN	798.088025

	key	Wire 1	Wire 2	Wire 3	Wire 4	Wire 5	Wire 6	Wire 7	Wire 8	Wire 9
0	1	2019-05-03 11:06:19	NaT	NaT	NaT	NaT	NaT	NaT	NaT	NaT
1	2	2019-05-03 11:36:50	NaT	NaT	NaT	NaT	NaT	NaT	NaT	NaT
2	3	2019-05-03 12:11:46	NaT	NaT	NaT	NaT	NaT	NaT	NaT	NaT
3	4	2019-05-03 12:43:22	NaT	NaT	NaT	NaT	NaT	NaT	NaT	NaT
4	5	2019-05-03 13:20:44	2019-05-03 13:15:34	NaT	NaT	NaT	NaT	NaT	NaT	NaT

	bulk_1	bulk_2	bulk_3	bulk_4	bulk_5	bulk_6	bulk_7	bulk_8	bulk_9	bulk_10	bulk_11	bulk_12	bulk_13	bulk_14	bulk_15	bulk_sum
key
1	0.0	0.0	0.0	43.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	206.0	0.0	150.0	154.0	553.0
2	0.0	0.0	0.0	73.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	206.0	0.0	149.0	154.0	582.0
3	0.0	0.0	0.0	34.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	205.0	0.0	152.0	153.0	544.0
4	0.0	0.0	0.0	81.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	207.0	0.0	153.0	154.0	595.0
5	0.0	0.0	0.0	78.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	203.0	0.0	151.0	152.0	584.0
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
3237	0.0	0.0	170.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	252.0	0.0	130.0	206.0	758.0
3238	0.0	0.0	126.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	254.0	0.0	108.0	106.0	594.0
3239	0.0	0.0	0.0	0.0	0.0	114.0	0.0	0.0	0.0	0.0	0.0	158.0	0.0	270.0	88.0	630.0
3240	0.0	0.0	0.0	0.0	0.0	26.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	192.0	54.0	272.0
3241	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	180.0	52.0	232.0

	Bulk 1	Bulk 2	Bulk 3	Bulk 4	Bulk 5	Bulk 6	Bulk 7	Bulk 8	Bulk 9	Bulk 10	Bulk 11	Bulk 12	Bulk 13	Bulk 14	Bulk 15
key
1	NaT	NaT	NaT	2019-05-03 11:28:48	NaT	NaT	NaT	NaT	NaT	NaT	NaT	2019-05-03 11:24:31	NaT	2019-05-03 11:14:50	2019-05-03 11:10:43
2	NaT	NaT	NaT	2019-05-03 11:36:50	NaT	NaT	NaT	NaT	NaT	NaT	NaT	2019-05-03 11:53:30	NaT	2019-05-03 11:48:37	2019-05-03 11:44:39
3	NaT	NaT	NaT	2019-05-03 12:32:39	NaT	NaT	NaT	NaT	NaT	NaT	NaT	2019-05-03 12:27:13	NaT	2019-05-03 12:21:01	2019-05-03 12:16:16
4	NaT	NaT	NaT	2019-05-03 12:43:22	NaT	NaT	NaT	NaT	NaT	NaT	NaT	2019-05-03 12:58:00	NaT	2019-05-03 12:51:11	2019-05-03 12:46:36
5	NaT	NaT	NaT	2019-05-03 13:30:47	NaT	NaT	NaT	NaT	NaT	NaT	NaT	2019-05-03 13:30:47	NaT	2019-05-03 13:34:12	2019-05-03 13:30:47
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
3237	NaT	NaT	2019-09-06 11:54:15	NaT	NaT	NaT	NaT	NaT	NaT	NaT	NaT	2019-09-06 11:49:45	NaT	2019-09-06 11:45:22	2019-09-06 11:40:06
3238	NaT	NaT	2019-09-06 12:26:52	NaT	NaT	NaT	NaT	NaT	NaT	NaT	NaT	2019-09-06 12:18:35	NaT	2019-09-06 12:31:49	2019-09-06 12:26:52
3239	NaT	NaT	NaT	NaT	NaT	2019-09-06 15:06:00	NaT	NaT	NaT	NaT	NaT	2019-09-06 15:01:44	NaT	2019-09-06 14:58:15	2019-09-06 14:48:06
3240	NaT	NaT	NaT	NaT	NaT	2019-09-06 16:24:28	NaT	NaT	NaT	NaT	NaT	NaT	NaT	2019-09-06 16:07:29	2019-09-06 16:01:34
3241	NaT	NaT	NaT	NaT	NaT	NaT	NaT	NaT	NaT	NaT	NaT	NaT	NaT	2019-09-06 17:26:33	2019-09-06 17:23:15

	кол-во замеров	кол-во партий	% кол-ва партий
0	5	892	27.736318
1	6	759	23.600746
2	4	520	16.169154
3	7	490	15.236318
4	8	205	6.374378
5	3	174	5.410448
6	9	84	2.611940
7	2	39	1.212687
8	10	28	0.870647
9	11	9	0.279851
10	12	5	0.155473
11	13	3	0.093284
12	1	2	0.062189
13	16	2	0.062189
14	14	2	0.062189
15	15	1	0.031095
16	17	1	0.031095

	Газ 1
count	3239.000000
mean	11.002062
std	6.220327
min	0.008399
25%	7.043089
50%	9.836267
75%	13.769915
max	77.995040

	key	Время замера	Температура
count	18092.000000	18092	14665.000000
mean	1616.460977	2019-07-05 13:36:58.791620608	1590.722741
min	1.000000	2019-05-03 11:02:04	1191.000000
25%	807.750000	2019-06-04 00:35:01.249999872	1580.000000
50%	1618.000000	2019-07-03 02:11:48	1590.000000
75%	2429.000000	2019-08-07 23:10:05.249999872	1599.000000
max	3241.000000	2019-09-06 17:30:05	1705.000000
std	934.641385	NaN	20.394381

	time_first_measurement	time_last_measurement	temperature_first_measurement	temperature_last_measurement	time_between_measurements
key
1	2019-05-03 11:02:04	2019-05-03 11:30:38	1571.0	1613.0	1714.0
2	2019-05-03 11:34:04	2019-05-03 11:55:09	1581.0	1602.0	1265.0
3	2019-05-03 12:06:44	2019-05-03 12:35:57	1596.0	1599.0	1753.0
4	2019-05-03 12:39:27	2019-05-03 12:59:47	1601.0	1625.0	1220.0
5	2019-05-03 13:11:03	2019-05-03 13:36:39	1576.0	1602.0	1536.0
...	...	...	...	...	...
2495	2019-08-10 11:27:47	2019-08-10 11:50:47	1570.0	1591.0	1380.0
2496	2019-08-10 11:56:48	2019-08-10 12:25:13	1554.0	1591.0	1705.0
2497	2019-08-10 12:37:26	2019-08-10 12:53:28	1571.0	1589.0	962.0
2498	2019-08-10 12:58:11	2019-08-10 13:23:31	1591.0	1594.0	1520.0
2499	2019-08-10 13:33:21	2019-08-10 13:58:58	1569.0	1603.0	1537.0

	time_first_measurement	time_last_measurement	temperature_first_measurement	temperature_last_measurement	time_between_measurements
key
867	2019-06-06 08:03:39	2019-06-06 08:48:23	1191.0	1599.0	2684.0
1214	2019-06-18 08:01:03	2019-06-18 08:43:56	1208.0	1591.0	2573.0
1619	2019-07-03 02:34:41	2019-07-03 02:43:59	1218.0	1590.0	558.0
2052	2019-07-25 08:49:15	2019-07-25 09:27:03	1227.0	1592.0	2268.0

	Wire 1	Wire 2	Wire 3	Wire 4	Wire 5	Wire 6	Wire 7	Wire 8	Wire 9
count	3055.000000	1079.000000	63.000000	14.000000	1.000	73.000000	11.000000	19.000000	29.000000
mean	100.895853	50.577323	189.482681	57.442841	15.132	48.016974	10.039007	53.625193	34.155752
std	42.012518	39.320216	99.513444	28.824667	NaN	33.919845	8.610584	16.881728	19.931616
min	1.918800	0.030160	0.144144	24.148801	15.132	0.034320	0.234208	45.076721	4.622800
25%	72.115684	20.193680	95.135044	40.807002	15.132	25.053600	6.762756	46.094879	22.058401
50%	100.158234	40.142956	235.194977	45.234282	15.132	42.076324	9.017009	46.279999	30.066399
75%	126.060483	70.227558	276.252014	76.124619	15.132	64.212723	11.886057	48.089603	43.862003
max	330.314424	282.780152	385.008668	113.231044	15.132	180.454575	32.847674	102.762401	90.053604

	wire_1	wire_2	wire_3	wire_4	wire_5	wire_6	wire_7	wire_8	wire_9	wire_sum
key
1	60.059998	0.00000	0.0	0.0	0.0	0.0	0.0	0.0	0.0	60.059998
2	96.052315	0.00000	0.0	0.0	0.0	0.0	0.0	0.0	0.0	96.052315
3	91.160157	0.00000	0.0	0.0	0.0	0.0	0.0	0.0	0.0	91.160157
4	89.063515	0.00000	0.0	0.0	0.0	0.0	0.0	0.0	0.0	89.063515
5	89.238236	9.11456	0.0	0.0	0.0	0.0	0.0	0.0	0.0	98.352796
...	...	...	...	...	...	...	...	...	...	...
3237	38.088959	0.00000	0.0	0.0	0.0	0.0	0.0	0.0	0.0	38.088959
3238	56.128799	0.00000	0.0	0.0	0.0	0.0	0.0	0.0	0.0	56.128799
3239	143.357761	0.00000	0.0	0.0	0.0	0.0	0.0	0.0	0.0	143.357761
3240	34.070400	0.00000	0.0	0.0	0.0	0.0	0.0	0.0	0.0	34.070400
3241	63.117595	0.00000	0.0	0.0	0.0	0.0	0.0	0.0	0.0	63.117595

	start_time_wire	finish_time_wire	duration_wire
count	3081	3081	3081.000000
mean	2019-07-05 19:30:10.739045632	2019-07-05 19:33:45.258357760	214.519312
min	2019-05-03 11:06:19	2019-05-03 11:06:19	0.000000
25%	2019-06-04 15:15:14	2019-06-04 15:15:14	0.000000
50%	2019-07-03 02:37:20	2019-07-03 02:41:39	0.000000
75%	2019-08-08 02:31:52	2019-08-08 02:31:52	401.000000
max	2019-09-06 17:10:06	2019-09-06 17:10:06	5937.000000
std	NaN	NaN	396.131967

	temperature_first_measurement	time_between_measurements	energy_consumption	bulk_1	bulk_3	bulk_4	bulk_5	bulk_6	bulk_10	bulk_11	...	wire_1	wire_2	wire_3	wire_6	wire_7	wire_9	wire_sum	duration_wire	bulk_9_wire_8	bulk_7_wire_4_bulk_2
key
1973	1562.0	1066.0	598.994020	0.0	310.0	0.0	0.0	93.0	0.0	0.0	...	40.098240	0.000000	0.000000	0.000000	0.0	0.0	40.098240	0.0	0.0	0.0
802	1580.0	3871.0	688.957621	0.0	190.0	0.0	0.0	46.0	0.0	0.0	...	29.249999	80.376400	0.000000	0.000000	0.0	0.0	109.626399	546.0	0.0	0.0
959	1586.0	8222.0	1229.980397	0.0	66.0	23.0	0.0	0.0	0.0	0.0	...	123.037201	0.000000	0.000000	0.000000	0.0	0.0	123.037201	0.0	0.0	0.0
680	1613.0	2993.0	578.532433	0.0	44.0	0.0	0.0	0.0	0.0	0.0	...	120.042000	0.000000	0.000000	0.000000	0.0	0.0	120.042000	0.0	0.0	0.0
970	1574.0	1498.0	219.296917	0.0	0.0	0.0	0.0	0.0	0.0	0.0	...	84.099601	0.000000	0.000000	0.000000	0.0	0.0	84.099601	0.0	0.0	0.0
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
119	1628.0	3897.0	200.972392	51.0	0.0	50.0	86.0	0.0	0.0	0.0	...	0.000000	0.000000	93.117027	43.174561	0.0	0.0	136.291588	1926.0	0.0	233.0
809	1587.0	2795.0	856.165815	0.0	405.0	0.0	0.0	0.0	0.0	0.0	...	137.638809	0.000000	0.000000	0.000000	0.0	0.0	137.638809	0.0	0.0	0.0
1775	1604.0	1744.0	351.696370	0.0	21.0	0.0	0.0	0.0	0.0	0.0	...	145.126801	0.000000	0.000000	0.000000	0.0	0.0	145.126801	0.0	0.0	0.0
1966	1527.0	2816.0	1112.650279	0.0	0.0	0.0	0.0	0.0	0.0	97.0	...	65.126877	0.000000	0.000000	0.000000	0.0	0.0	65.126877	0.0	0.0	0.0
1580	1579.0	1327.0	595.244297	0.0	0.0	0.0	0.0	0.0	0.0	0.0	...	13.993200	60.078716	0.000000	0.000000	0.0	0.0	74.071916	450.0	0.0	0.0

	temperature_first_measurement	time_between_measurements	energy_consumption	bulk_1	bulk_3	bulk_4	bulk_5	bulk_6	bulk_10	bulk_11	...	wire_1	wire_2	wire_3	wire_6	wire_7	wire_9	wire_sum	duration_wire	bulk_9_wire_8	bulk_7_wire_4_bulk_2
key
1480	1567.0	1255.0	443.225773	0.0	0.0	107.0	0.0	0.0	0.0	0.0	...	122.990398	0.000000	0.0	0.0	0.0	0.0	122.990398	0.0	0.0	0.0
2322	1616.0	3225.0	793.032361	52.0	154.0	0.0	0.0	0.0	0.0	0.0	...	172.273911	0.000000	0.0	0.0	0.0	0.0	172.273911	0.0	0.0	0.0
673	1609.0	2671.0	715.322078	0.0	87.0	0.0	0.0	0.0	0.0	0.0	...	134.877605	0.000000	0.0	0.0	0.0	0.0	134.877605	0.0	0.0	0.0
660	1619.0	3440.0	768.955278	0.0	0.0	110.0	0.0	0.0	0.0	0.0	...	76.143600	0.000000	0.0	0.0	0.0	0.0	76.143600	0.0	0.0	0.0
1357	1538.0	1989.0	1046.797454	0.0	0.0	231.0	0.0	0.0	0.0	0.0	...	105.112795	0.000000	0.0	0.0	0.0	0.0	105.112795	0.0	0.0	0.0
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
335	1564.0	4258.0	1556.605508	0.0	0.0	0.0	0.0	162.0	0.0	0.0	...	73.429200	0.000000	0.0	0.0	0.0	0.0	73.429200	0.0	0.0	0.0
2151	1622.0	7772.0	1257.999589	0.0	0.0	0.0	0.0	0.0	126.0	0.0	...	145.236007	0.000000	0.0	0.0	0.0	0.0	145.236007	0.0	0.0	0.0
537	1590.0	4248.0	974.347090	0.0	0.0	0.0	0.0	0.0	0.0	0.0	...	100.011597	0.000000	0.0	0.0	0.0	0.0	100.011597	0.0	0.0	0.0
650	1577.0	6106.0	1885.358882	0.0	0.0	19.0	0.0	0.0	0.0	0.0	...	170.398798	0.000000	0.0	0.0	0.0	0.0	170.398798	0.0	0.0	0.0
569	1599.0	1352.0	862.814124	0.0	290.0	0.0	0.0	0.0	0.0	0.0	...	49.186798	60.169201	0.0	0.0	0.0	0.0	109.355999	477.0	0.0	0.0

	temperature_first_measurement	time_between_measurements	energy_consumption	bulk_1	bulk_3	bulk_4	bulk_5	bulk_6	bulk_10	bulk_11	...	wire_1	wire_2	wire_3	wire_6	wire_7	wire_9	wire_sum	duration_wire	bulk_9_wire_8	bulk_7_wire_4_bulk_2
key
1973	-1.068818	-0.869244	-0.222219	-0.278179	3.468995	-0.627861	-0.125004	1.221972	-0.243976	-0.186131	...	-1.431246	-0.524372	-0.124387	-0.120551	-0.035756	-0.091804	-1.321379	-0.522456	-0.072593	-0.076135
802	-0.304182	1.037816	0.057718	-0.278179	1.879286	-0.627861	-0.125004	0.404780	-0.243976	-0.186131	...	-1.680569	1.880855	-0.124387	-0.120551	-0.035756	-0.091804	-0.245646	0.794695	-0.072593	-0.076135
959	-0.049304	3.995967	1.741205	-0.278179	0.236587	-0.238342	-0.125004	-0.395024	-0.243976	-0.186131	...	0.474928	-0.524372	-0.124387	-0.120551	-0.035756	-0.091804	-0.038155	-0.522456	-0.072593	-0.076135
680	1.097649	0.440882	-0.285889	-0.278179	-0.054860	-0.627861	-0.125004	-0.395024	-0.243976	-0.186131	...	0.406090	-0.524372	-0.124387	-0.120551	-0.035756	-0.091804	-0.084496	-0.522456	-0.072593	-0.076135
970	-0.559061	-0.575536	-1.403713	-0.278179	-0.637753	-0.627861	-0.125004	-0.395024	-0.243976	-0.186131	...	-0.419969	-0.524372	-0.124387	-0.120551	-0.035756	-0.091804	-0.640594	-0.522456	-0.072593	-0.076135
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
119	1.734846	1.055492	-1.460733	4.077678	-0.637753	0.218919	3.494437	-0.395024	-0.243976	-0.186131	...	-2.352818	-0.524372	3.085229	4.644601	-0.035756	-0.091804	0.166916	4.123757	-0.072593	4.430674
809	-0.006824	0.306266	0.578016	-0.278179	4.727515	-0.627861	-0.125004	-0.395024	-0.243976	-0.186131	...	0.810515	-0.524372	-0.124387	-0.120551	-0.035756	-0.091804	0.187760	-0.522456	-0.072593	-0.076135
1775	0.715332	-0.408286	-0.991729	-0.278179	-0.359554	-0.627861	-0.125004	-0.395024	-0.243976	-0.186131	...	0.982610	-0.524372	-0.124387	-0.120551	-0.035756	-0.091804	0.303614	-0.522456	-0.072593	-0.076135
1966	-2.555609	0.320544	1.376112	-0.278179	-0.637753	-0.627861	-0.125004	-0.395024	-0.243976	3.968230	...	-0.856016	-0.524372	-0.124387	-0.120551	-0.035756	-0.091804	-0.934138	-0.522456	-0.072593	-0.076135
1580	-0.346662	-0.691795	-0.233887	-0.278179	-0.637753	-0.627861	-0.125004	-0.395024	-0.243976	-0.186131	...	-2.031214	1.273456	-0.124387	-0.120551	-0.035756	-0.091804	-0.795742	0.563108	-0.072593	-0.076135

Содержание.

Библиотеки и константы.¶

Знакомство с данными.¶

Предобработка данных.¶

df_arc — данные об электродах.¶

df_bulk — данные о подаче сыпучих материалов (объём)¶

df_bulk_time — данные о подаче сыпучих материалов (время).¶

df_gas — данные о продувке сплава газом¶

df_temp — результаты измерения температуры.¶

df_wire — данные о проволочных материалах (объём).¶

df_wire_time — данные о проволочных материалах (время).¶

Моменты, которые надо уточнить и согласовать с представителем заказчика.¶

Результат предобработки.¶

Объединение таблиц.¶

Обучение моделей.¶

Подготовка выборок.¶

Вычислим среднюю абсолютную ошибку (MAE) для модели, которая всегда предсказывает среднее значение целевой переменной.¶

Линейная регрессия.¶

Случайный лес RandomForestRegressor.¶

Градиентный бустинг LightGBM¶

Градиентный бустинг CatBoostRegressor.¶

Support Vector Regression.¶

Проанализируем качество моделей.¶

Тестирование лучшей модели.¶

Степень влияния признаков на целевой показатель.¶

Отчёт по решению.¶

`df_arc` — данные об электродах.¶

`df_bulk` — данные о подаче сыпучих материалов (объём)¶

`df_bulk_time` — данные о подаче сыпучих материалов (время).¶

`df_gas` — данные о продувке сплава газом¶

`df_temp` — результаты измерения температуры.¶

`df_wire` — данные о проволочных материалах (объём).¶

`df_wire_time` — данные о проволочных материалах (время).¶

	temperature_first_measurement	time_between_measurements	energy_consumption	bulk_1	bulk_3	bulk_4	bulk_5	bulk_6	bulk_10	bulk_11	...	wire_1	wire_2	wire_3	wire_6	wire_7	wire_9	wire_sum	duration_wire	bulk_9_wire_8	bulk_7_wire_4_bulk_2
key
1480	-0.856419	-0.740747	-0.706919	-0.278179	-0.637753	1.184248	-0.125004	-0.395024	-0.243976	-0.186131	...	0.473853	-0.524372	-0.124387	-0.120551	-0.035756	-0.091804	-0.038879	-0.522456	-0.072593	-0.076135
2322	1.225089	0.598614	0.381565	4.163087	1.402373	-0.627861	-0.125004	-0.395024	-0.243976	-0.186131	...	1.606528	-0.524372	-0.124387	-0.120551	-0.035756	-0.091804	0.723631	-0.522456	-0.072593	-0.076135
673	0.927730	0.221961	0.139756	-0.278179	0.514786	-0.627861	-0.125004	-0.395024	-0.243976	-0.186131	...	0.747054	-0.524372	-0.124387	-0.120551	-0.035756	-0.091804	0.145039	-0.522456	-0.072593	-0.076135
660	1.352528	0.744788	0.306645	-0.278179	-0.637753	1.235055	-0.125004	-0.395024	-0.243976	-0.186131	...	-0.602821	-0.524372	-0.124387	-0.120551	-0.035756	-0.091804	-0.763689	-0.522456	-0.072593	-0.076135
1357	-2.088332	-0.241716	1.171199	-0.278179	-0.637753	3.284263	-0.125004	-0.395024	-0.243976	-0.186131	...	0.062974	-0.524372	-0.124387	-0.120551	-0.035756	-0.091804	-0.315480	-0.522456	-0.072593	-0.076135
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
335	-0.983858	1.300929	2.757556	-0.278179	-0.637753	-0.627861	-0.125004	2.421678	-0.243976	-0.186131	...	-0.665205	-0.524372	-0.124387	-0.120551	-0.035756	-0.091804	-0.805686	-0.522456	-0.072593	-0.076135
2151	1.479967	3.690021	1.828392	-0.278179	-0.637753	-0.627861	-0.125004	-0.395024	5.645407	-0.186131	...	0.985120	-0.524372	-0.124387	-0.120551	-0.035756	-0.091804	0.305303	-0.522456	-0.072593	-0.076135
537	0.120615	1.294130	0.945757	-0.278179	-0.637753	-0.627861	-0.125004	-0.395024	-0.243976	-0.186131	...	-0.054266	-0.524372	-0.124387	-0.120551	-0.035756	-0.091804	-0.394405	-0.522456	-0.072593	-0.076135
650	-0.431622	2.557344	3.780529	-0.278179	-0.637753	-0.306085	-0.125004	-0.395024	-0.243976	-0.186131	...	1.563433	-0.524372	-0.124387	-0.120551	-0.035756	-0.091804	0.694620	-0.522456	-0.072593	-0.076135
569	0.502933	-0.674798	0.598703	-0.278179	3.204044	-0.627861	-0.125004	-0.395024	-0.243976	-0.186131	...	-1.222365	1.276164	-0.124387	-0.120551	-0.035756	-0.091804	-0.249829	0.628242	-0.072593	-0.076135

	n_estimators	max_depth	min_samples_split	min_samples_leaf	max_features	bootstrap	score
0	102.0	18.0	6.0	3.0	None	1.0	6.43
1	102.0	11.0	5.0	3.0	None	1.0	6.43
2	107.0	27.0	5.0	3.0	None	1.0	6.43

	n_estimators	max_depth	min_samples_split	min_samples_leaf	max_features	bootstrap	score
0	85.0	12.0	4.0	3.0	None	1.0	6.43
1	111.0	12.0	6.0	3.0	None	1.0	6.43
2	106.0	11.0	6.0	3.0	None	1.0	6.43
3	54.0	12.0	5.0	2.0	None	1.0	6.43
4	54.0	12.0	5.0	2.0	None	1.0	6.43
5	92.0	17.0	3.0	3.0	None	1.0	6.43

	n_estimators	max_depth	min_samples_split	min_samples_leaf	max_features	bootstrap	score
0	272.0	14.0	6.0	2.0	None	1.0	6.43
43	291.0	14.0	6.0	2.0	None	1.0	6.43
31	257.0	14.0	6.0	2.0	None	1.0	6.43
32	282.0	14.0	6.0	2.0	None	1.0	6.43
33	259.0	14.0	6.0	2.0	None	1.0	6.43
34	287.0	14.0	6.0	2.0	None	1.0	6.43
35	260.0	14.0	6.0	2.0	None	1.0	6.43
36	254.0	14.0	6.0	2.0	None	1.0	6.43
37	280.0	14.0	6.0	2.0	None	1.0	6.43
38	275.0	14.0	6.0	2.0	None	1.0	6.43

	n_estimators	max_depth	min_child_samples	learning_rate	subsample	colsample_bytree	score
0	276.0	2.0	7.0	0.09	0.76	0.7	6.25
16	274.0	2.0	7.0	0.09	0.67	0.7	6.25
29	264.0	2.0	7.0	0.09	0.67	0.7	6.25
28	281.0	2.0	7.0	0.09	0.69	0.7	6.25
27	276.0	2.0	7.0	0.09	0.67	0.7	6.25
26	279.0	2.0	7.0	0.09	0.68	0.7	6.25
25	278.0	2.0	7.0	0.09	0.67	0.7	6.25
24	276.0	2.0	7.0	0.09	0.67	0.7	6.25
23	281.0	2.0	7.0	0.09	0.69	0.7	6.25
22	276.0	2.0	7.0	0.09	0.68	0.7	6.25

	iterations	depth	learning_rate	score
0	278.0	4.0	0.08	6.29
1	346.0	5.0	0.06	6.29
2	305.0	5.0	0.07	6.29
3	329.0	6.0	0.04	6.29

	iterations	depth	learning_rate	score
0	373.0	6.0	0.04	6.28
58	358.0	6.0	0.04	6.28
67	363.0	6.0	0.04	6.28
66	359.0	6.0	0.04	6.28
65	350.0	6.0	0.04	6.28
64	354.0	6.0	0.04	6.28
63	368.0	6.0	0.04	6.28
62	336.0	6.0	0.04	6.28
61	361.0	6.0	0.04	6.28
60	364.0	6.0	0.04	6.28

	Признаки	Важность (%)
0	energy_consumption	12.816300
1	temperature_first_measurement	12.421952
2	time_between_measurements	9.431482
3	bulk_14	8.839961
4	wire_1	8.149852
5	gas_quantities	8.149852
6	duration_bulk	6.539599
7	bulk_12	5.619454
8	wire_sum	5.455143
9	bulk_6	4.042064
10	wire_2	3.746303
11	bulk_15	3.450542
12	bulk_3	3.220506
13	duration_wire	2.136050
14	bulk_1	1.840289
15	bulk_4	1.675978
16	bulk_11	0.953007
17	bulk_10	0.525797
18	wire_6	0.492935
19	bulk_5	0.262898
20	bulk_7_wire_4_bulk_2	0.131449
21	wire_3	0.098587
22	wire_7	0.000000
23	bulk_13	0.000000
24	bulk_9_wire_8	0.000000
25	wire_9	0.000000